From Assuptions to Evidence - An Evidence-Based Framework for Digital Product Discovery in the Age of Artificial Intelligence

Abstract

Digital product development has become increasingly efficient through advances in agile methodologies, cloud computing and, more recently, generative artificial intelligence (AI). While these developments have dramatically reduced the cost and time required to develop software, evidence suggests that high failure rates of digital products persist. Existing literature indicates that many failures arise not from technical shortcomings but from inadequate problem definition, weak customer understanding and premature commitment to product solutions. Consequently, organisations frequently optimise software delivery while underinvesting in product discovery and evidence generation.

This paper introduces the HDNA Framework as a conceptual model for evidence-based digital product discovery. Drawing upon research from innovation management, Lean Startup, design thinking, human-centred design, product management and organisational learning, the framework proposes a structured sequence of discovery activities designed to progressively reduce uncertainty before significant development investment occurs. Unlike approaches primarily focused on software delivery, HDNA integrates strategic framing, customer understanding, hypothesis validation and iterative decision loops into a coherent decision-making architecture.

The paper further examines how generative AI fundamentally changes software economics by dramatically lowering development costs while simultaneously increasing the strategic importance of problem discovery. As software production becomes increasingly commoditised, competitive advantage shifts towards organisations capable of identifying meaningful customer problems and systematically validating opportunities before scaling implementation.

The article contributes to digital innovation literature by synthesising existing research into an integrated conceptual framework that connects strategic management, customer-centric innovation and continuous evidence generation. It concludes by discussing implications for practitioners and identifies directions for future empirical validation of the HDNA framework.

Keywords

Digital Product Development; Product Discovery; Innovation Management; Human-Centred Design; Evidence-Based Management; Lean Startup; Artificial Intelligence; Product Strategy; Customer Discovery; Digital Innovation

1. Introduction

Digital technologies have fundamentally transformed how organisations innovate, compete and create value. Over the past two decades, advances in cloud computing, application programming interfaces (APIs), platform ecosystems, DevOps, continuous integration and continuous deployment (CI/CD) have dramatically reduced the technical barriers to software development. More recently, the emergence of generative artificial intelligence (AI) has accelerated this transformation by automating activities that previously required substantial human expertise, including software coding, interface design, documentation, testing and requirements generation (Noy & Zhang, 2023; Brynjolfsson, Li & Raymond, 2025). Collectively, these developments have shifted software engineering from a resource-intensive activity towards an increasingly automated capability, enabling organisations to build sophisticated digital products with unprecedented speed and efficiency.

Despite these technological advances, evidence consistently demonstrates that improving software development capability alone has not translated into higher rates of digital product success. While organisations have become increasingly proficient at delivering software rapidly, the commercial performance of new digital products remains highly uncertain. Studies of new product development continue to report substantial failure rates across industries, with estimates frequently suggesting that between 35 and 60 per cent of new products fail to achieve their commercial objectives, despite significant investment in research, development and marketing (Castellion & Markham, 2013; Cooper, 2019). Within digital environments, where product development cycles are shorter and competitive pressures are greater, failure often occurs not because organisations lack technical capability but because they misidentify customer problems, overestimate market demand or prematurely commit resources to poorly validated solutions.

This apparent contradiction exposes an important paradox within contemporary digital innovation. Organisations have invested heavily in improving software delivery through Agile development, DevOps and platform engineering, yet comparatively less attention has been devoted to improving the quality of strategic decisions that precede implementation. Agile methodologies have transformed software engineering by enabling iterative development, continuous customer feedback and adaptive planning (Beck et al., 2001; Dingsøyr et al., 2012). Likewise, DevOps practices have substantially improved deployment frequency, software reliability and organisational responsiveness (Forsgren, Humble & Kim, 2018). However, these approaches primarily optimise how software is developed rather than addressing the more fundamental strategic question of whether the proposed solution should be developed in the first place.

This distinction between delivery efficiency and discovery effectiveness has become increasingly significant within both product management research and innovation management. Whereas delivery focuses on implementing validated ideas efficiently, product discovery concerns reducing uncertainty before major development investment occurs by systematically investigating customer problems, market opportunities and organisational capabilities (Blank, 2013; Torres, 2021). Contemporary research increasingly argues that organisations often excel at building products while remaining comparatively weak at identifying opportunities that genuinely create customer value (Perri, 2018). Consequently, many firms become highly efficient feature factories, optimising development throughput while simultaneously reducing their ability to challenge underlying strategic assumptions.

Underlying this challenge is a broader problem of organisational decision-making under conditions of uncertainty. Innovation differs fundamentally from routine operational work because decisions must be made in situations characterised by incomplete information, ambiguous customer needs and rapidly changing market conditions (March, 1991; Teece, 2018). Rather than executing predetermined plans, innovation requires organisations to continuously generate, test and revise hypotheses regarding customer behaviour, competitive dynamics and technological change. Failure therefore frequently reflects weaknesses in organisational learning rather than deficiencies in engineering capability. As Argyris and Schön (1978) argue, organisations often improve existing routines without questioning the assumptions upon which those routines are based, creating conditions in which ineffective strategic decisions become progressively embedded within organisational processes.

These challenges are amplified by well-established cognitive biases that influence managerial decision-making. Confirmation bias encourages decision-makers to seek evidence supporting existing beliefs while discounting contradictory information (Kahneman, 2011). Escalation of commitment further increases the likelihood that organisations continue investing in failing initiatives because previous investments create psychological, political and financial pressures to persist (Staw, 1976). Groupthink and organisational inertia may similarly inhibit critical reflection, particularly where product roadmaps become politically committed before customer evidence has been systematically gathered. Consequently, many organisations undertake customer research only after significant strategic decisions have already been made, reducing research to a confirmatory exercise rather than a mechanism for genuine discovery (Liedtka, 2018).

The emergence of generative AI fundamentally changes the economics of this problem. Large language models and AI-assisted software engineering tools have dramatically reduced the marginal cost of creating software artefacts, enabling organisations to prototype, develop and deploy digital products at speeds previously unattainable (Brynjolfsson, Li & Raymond, 2025). Early empirical evidence suggests significant productivity improvements among knowledge workers using generative AI, particularly in programming, documentation and analytical tasks (Noy & Zhang, 2023). From a technological perspective, this represents one of the most significant advances in software development since the introduction of Agile methodologies.

Paradoxically, however, reducing the cost of software production does not reduce uncertainty regarding customer value. AI can generate thousands of product ideas, functional prototypes and production-ready code, but it cannot independently determine whether customers experience a sufficiently important problem to justify a proposed solution, nor can it reliably assess whether a product aligns with an organisation's strategic capabilities or competitive positioning. In economic terms, AI reduces the scarcity of software production while increasing the relative scarcity of customer insight, strategic judgement and organisational learning. As Simon (1971) argued, an abundance of information shifts scarcity towards attention; similarly, an abundance of software generation shifts competitive advantage towards selecting the right opportunities rather than merely executing them efficiently.

This changing economic landscape suggests that the strategic bottleneck in digital innovation has shifted. Historically, organisations competed through superior engineering capability because software development represented the principal organisational constraint. Increasingly, however, competitive advantage depends upon an organisation's ability to identify valuable customer problems, generate reliable evidence and make informed investment decisions before committing significant development resources. Within the resource-based view of the firm, sustainable competitive advantage derives from valuable, rare and difficult-to-imitate organisational capabilities rather than from widely accessible technologies (Barney, 1991). As AI progressively commoditises software engineering, capabilities relating to customer understanding, evidence generation and organisational learning become correspondingly more valuable.

Although several influential approaches—including Lean Startup (Ries, 2011), Customer Development (Blank, 2013), Design Thinking (Brown, 2009), Human-Centred Design (ISO, 2019), Jobs-to-be-Done theory (Christensen et al., 2016) and Continuous Discovery (Torres, 2021)—have advanced understanding of product discovery, they typically address particular aspects of innovation rather than providing a comprehensive decision architecture that integrates strategic framing, customer understanding, hypothesis generation, evidence evaluation and organisational learning into a coherent process. Existing methodologies frequently operate independently, resulting in fragmented discovery activities that may generate valuable insights but remain weakly connected to organisational governance and strategic decision-making. Consequently, organisations often adopt multiple complementary methodologies without establishing a systematic mechanism through which evidence accumulates, informs investment decisions and continuously reshapes strategic direction.

This fragmentation represents an important gap within the literature. While considerable research has examined customer discovery, Lean experimentation, human-centred design and evidence-based management individually, comparatively little attention has been devoted to integrating these perspectives into a unified framework capable of supporting evidence-based decision-making throughout digital product discovery. Moreover, despite rapidly expanding research into generative AI, relatively few studies have examined how AI fundamentally changes the relative importance of discovery and delivery within digital innovation. Much of the emerging literature focuses on AI as a productivity-enhancing technology, whereas its broader implications for product strategy, organisational learning and innovation governance remain underexplored.

Accordingly, this paper argues that the principal challenge facing contemporary digital product development is no longer simply improving software delivery but improving the quality of strategic learning that precedes implementation. It proposes that product discovery should be understood as a systematic process of transforming assumptions into evidence through iterative experimentation, customer engagement and organisational learning. Building upon established theories of innovation management, organisational learning, evidence-based management and product discovery, the paper develops the HDNA Framework as an integrated conceptual model designed to reduce uncertainty before substantial implementation commitments are made.

1.1 Research Aim and Research Questions

Against this backdrop, the central aim of this paper is to develop an integrated conceptual framework for evidence-based digital product discovery that enables organisations to systematically reduce uncertainty before committing significant resources to product development. Rather than proposing another standalone innovation methodology, the paper synthesises established theories from innovation management, product management, organisational learning, human-centred design and evidence-based management into a coherent decision architecture. The resulting HDNA Framework is intended to support organisations in replacing intuition, implicit assumptions and political decision-making with progressively stronger empirical evidence throughout the discovery process.

The framework is founded on the proposition that uncertainty is not a problem to be eliminated at the outset of innovation but a condition to be managed through structured learning. Consequently, the primary output of early-stage product development is not software, prototypes or product specifications, but validated knowledge regarding customer problems, market opportunities and organisational capabilities. This perspective extends evidence-based management (Rousseau, 2006) into the domain of digital product innovation by treating every significant product decision as a hypothesis requiring empirical investigation rather than executive consensus.

To address the research problem identified above, the paper is guided by three interrelated research questions.

RQ1: Why do digital products continue to fail despite substantial improvements in software development capabilities?

This question addresses the apparent contradiction between significant advances in software engineering and the persistent failure of many digital products to achieve commercial success. While Agile methodologies, DevOps practices and AI-assisted software development have transformed implementation capability, evidence suggests that product failure is more frequently associated with inadequate customer understanding, weak problem definition and poor strategic alignment than with technical deficiencies (Cooper, 2019; Castellion & Markham, 2013). Investigating this question contributes to a more nuanced understanding of where uncertainty originates during digital innovation and why improvements in delivery efficiency alone have proved insufficient.

RQ2: How has generative artificial intelligence changed the relative importance of product discovery compared with software delivery?

This question reflects the growing recognition that generative AI fundamentally alters the economics of software development. Existing research has predominantly examined AI through the lens of productivity enhancement, focusing on code generation, documentation and automation (Noy & Zhang, 2023; Brynjolfsson, Li & Raymond, 2025). Comparatively little attention has been devoted to understanding how AI reshapes the strategic balance between implementation and discovery. This paper argues that as software creation becomes increasingly commoditised, the scarcity underpinning competitive advantage shifts towards customer insight, strategic judgement and organisational learning. Accordingly, the paper explores AI not merely as an operational technology but as a catalyst that redefines where organisations create value during digital innovation.

RQ3: How can organisations structure evidence generation throughout product discovery to improve strategic decision quality?

The third research question seeks to address the fragmented nature of existing discovery methodologies. Although organisations increasingly adopt practices such as Design Thinking, Lean Startup, Customer Development and Continuous Discovery, these approaches are frequently implemented independently, creating discontinuities between customer research, strategic planning, governance and product delivery. This paper therefore investigates how evidence generation can be organised into a coherent learning system that progressively reduces uncertainty across the entire discovery lifecycle. The HDNA Framework represents a conceptual response to this question by integrating strategic framing, customer understanding, hypothesis validation and iterative learning within a unified decision architecture.

Collectively, these research questions shift attention from software production towards organisational learning. Rather than asking how organisations can build software more efficiently, they examine how organisations can make better strategic decisions before software is built. This distinction reflects an emerging consensus within innovation management that sustainable competitive advantage increasingly depends on the quality of organisational learning rather than solely on operational efficiency (March, 1991; Teece, 2018).

1.2 Theoretical Contributions

This paper makes several contributions to the literature on digital innovation and product management.

First, it contributes to the growing body of research advocating evidence-based approaches to innovation by integrating previously fragmented theoretical perspectives into a single conceptual framework. Existing methodologies—including Agile software development, Lean Startup, Customer Development, Human-Centred Design, Jobs-to-be-Done theory and Continuous Discovery—each address important dimensions of innovation but typically focus on specific phases or activities within the product development lifecycle. By synthesising these complementary perspectives, the HDNA Framework offers a broader conceptual architecture that explicitly links strategic intent, customer understanding, experimentation and governance through continuous evidence generation.

Second, the paper extends existing discussions concerning the strategic implications of generative AI. Much of the current literature examines AI primarily as a productivity-enhancing technology capable of improving programming efficiency and reducing development costs. While these contributions are significant, they provide only a partial account of AI's broader organisational implications. This paper argues that AI fundamentally changes the locus of competitive advantage by reducing the scarcity of software engineering while increasing the strategic importance of opportunity identification, customer understanding and evidence-based decision-making. In doing so, it positions AI as a force that elevates, rather than diminishes, the importance of product discovery.

Third, the paper contributes to organisational learning theory by conceptualising product discovery as a continuous learning capability rather than a discrete project phase. Drawing upon the work of Argyris and Schön (1978), March (1991) and Teece (2018), the framework treats innovation as an ongoing process of organisational adaptation in which assumptions are continuously tested, revised or abandoned in response to empirical evidence. Progress is therefore measured not by the volume of software produced but by the quality of knowledge generated. This reframing connects contemporary product management with broader theories of knowledge creation, dynamic capabilities and organisational adaptation, providing a richer theoretical foundation for future empirical investigation.

1.3 Practical Contributions

Beyond its theoretical significance, the paper offers several practical contributions for managers responsible for digital innovation.

The HDNA Framework provides organisations with a structured approach for governing product discovery through evidence rather than intuition. By explicitly documenting assumptions, formulating hypotheses and evaluating evidence before increasing investment, organisations can reduce the likelihood of committing substantial resources to products that lack meaningful customer demand. This approach complements rather than replaces existing delivery methodologies, strengthening the strategic decisions that precede implementation while allowing Agile and DevOps practices to remain focused on efficient execution.

The framework also provides a practical mechanism for integrating customer research with organisational strategy. Rather than treating discovery as an isolated design activity, HDNA positions customer evidence within a broader strategic context that includes organisational capabilities, competitive dynamics and technological change. Consequently, product discovery becomes a governance capability embedded within organisational decision-making rather than an optional precursor to software development.

Finally, the framework offers a foundation for organisations seeking to incorporate AI into product management responsibly. While AI can accelerate coding, prototyping and analysis, the framework emphasises that strategic decisions continue to require human judgement, contextual understanding and empirical validation. AI therefore functions as a partner in organisational learning rather than a substitute for managerial decision-making.

1.4 Structure of the Paper

The remainder of this paper is organised as follows. Section 2 reviews the literature underpinning evidence-based product discovery, drawing upon research in innovation management, Agile software development, Lean Startup, Customer Development, Human-Centred Design, Jobs-to-be-Done theory, organisational learning and generative artificial intelligence. Particular attention is given to identifying areas of theoretical convergence and fragmentation that motivate the development of an integrated framework.

Section 3 introduces the HDNA Framework, describing its conceptual foundations, four iterative discovery phases and evidence-governed learning loops. The framework is positioned relative to existing innovation methodologies, and a series of theoretical propositions is developed to guide future empirical research.

Section 4 discusses the broader implications of the framework within the context of AI-enabled digital innovation. It argues that declining software production costs increase the strategic importance of customer understanding, organisational learning and evidence generation, reframing product discovery as a core source of competitive advantage.

Section 5 examines the managerial implications of adopting evidence-based product discovery, including changes to governance, performance measurement, product leadership and organisational capability development. It also proposes a maturity model through which organisations can assess their progress towards evidence-governed innovation.

Finally, Section 6 concludes by revisiting the research questions, summarising the theoretical and practical contributions of the paper, acknowledging its limitations and outlining opportunities for future empirical investigation.

1.5 Chapter Summary

Digital product development is entering a period in which the principal constraint on innovation is no longer the ability to build software but the ability to determine what should be built, for whom and why. Although advances in Agile development, cloud computing and generative AI have dramatically improved implementation capability, they have not resolved the fundamental uncertainty that characterises innovation. Organisations continue to fail not because they cannot engineer sophisticated products but because they often lack robust evidence that those products address meaningful customer problems within strategically attractive markets.

This paper contends that the strategic challenge of digital innovation has therefore shifted from software delivery to evidence generation. As AI progressively reduces the cost of implementation, customer insight, organisational learning and strategic judgement become increasingly scarce sources of competitive advantage. Building on this premise, the paper develops the HDNA Framework as an integrated model for evidence-based product discovery that synthesises established theories across innovation management, organisational learning and product management into a coherent decision architecture. By conceptualising product discovery as a cumulative process of transforming assumptions into validated knowledge, the framework seeks to reposition evidence—not software—as the primary output of early-stage innovation. The following chapter reviews the theoretical foundations upon which this framework is constructed.

2. Literature Review

2.1 Digital Innovation as a Process of Uncertainty Reduction

Digital innovation differs fundamentally from traditional product development because it occurs under conditions characterised by technological uncertainty, evolving customer needs and rapidly changing competitive environments. Unlike mature engineering disciplines, where design requirements can often be specified with reasonable certainty before implementation, digital innovation is inherently exploratory. Organisations rarely possess complete knowledge of customer problems, market demand or technological feasibility at the outset of product development. Consequently, innovation should be understood not as the execution of predetermined plans but as a process of progressively reducing uncertainty through organisational learning and evidence generation (March, 1991; Teece, 2018).

This distinction has become increasingly important as software-intensive products have come to dominate contemporary economies. Digital products exhibit characteristics that differentiate them from traditional physical products, including low marginal production costs, rapid iteration cycles, continuous post-launch evolution and strong network effects (Yoo et al., 2010; Nambisan, Lyytinen, Majchrzak & Song, 2017). These characteristics create opportunities for continuous experimentation but simultaneously increase uncertainty because customer expectations, technologies and competitive landscapes evolve throughout the product lifecycle. Rather than delivering a finished artefact, organisations increasingly manage products as continuously evolving services that require ongoing adaptation.

Innovation scholars have long argued that uncertainty constitutes the defining characteristic of innovation. March's (1991) distinction between exploration and exploitation remains particularly influential. Exploration encompasses experimentation, discovery and the pursuit of new knowledge, whereas exploitation focuses on refinement, efficiency and execution. Successful organisations must balance these competing activities, recognising that excessive emphasis on exploitation may improve operational performance while simultaneously reducing an organisation's capacity to identify emerging opportunities. In digital product development, this balance is reflected in the relationship between product discovery and product delivery. Delivery primarily represents exploitation through efficient implementation, whereas discovery embodies exploration by generating evidence regarding customer needs, market opportunities and strategic alternatives.

Organisational learning theory similarly conceptualises innovation as an iterative process of knowledge creation rather than linear execution. Argyris and Schön (1978) distinguish between single-loop learning, in which organisations improve existing practices without questioning underlying assumptions, and double-loop learning, where fundamental beliefs, goals and mental models are critically examined and revised. Digital product discovery frequently requires double-loop learning because initial assumptions regarding customer behaviour, market demand and product value are often incomplete or incorrect. Organisations that fail to challenge these assumptions risk embedding strategic errors into product roadmaps, technical architectures and investment decisions, making subsequent adaptation increasingly costly.

Knowledge creation provides another theoretical lens through which to understand digital innovation. Nonaka and Takeuchi (1995) argue that organisational innovation depends upon the continuous interaction between tacit and explicit knowledge through processes of socialisation, externalisation, combination and internalisation. Within product discovery, tacit knowledge derived from customer interactions, stakeholder experience and market observation is progressively transformed into explicit hypotheses, prototypes, experiments and validated evidence. Innovation therefore emerges not simply from technological capability but from an organisation's ability to convert dispersed knowledge into informed strategic decisions.

These perspectives collectively suggest that uncertainty should not be viewed as a temporary obstacle to innovation but as its defining condition. The objective of product discovery is therefore not to eliminate uncertainty entirely—an impossible objective within dynamic markets—but to reduce critical uncertainties sufficiently to support informed investment decisions. Evidence generation becomes the mechanism through which organisations progressively replace assumptions with knowledge, enabling decisions to evolve alongside emerging insights rather than remaining constrained by initial beliefs.

This interpretation closely aligns with evidence-based management, which advocates that managerial decisions should be informed by the best available evidence rather than intuition, organisational tradition or hierarchical authority (Rousseau, 2006; Pfeffer & Sutton, 2006). Although evidence-based management has been widely discussed within organisational science, its application to digital product development remains comparatively underdeveloped. Existing product management methodologies frequently encourage experimentation but provide less guidance regarding how evidence should accumulate across successive discovery activities or inform governance decisions at the organisational level. This represents an important theoretical gap that motivates the present research.

Furthermore, behavioural research demonstrates that uncertainty is frequently exacerbated by systematic cognitive biases. Kahneman (2011) highlights how confirmation bias, overconfidence and availability heuristics influence judgement under uncertainty, while Staw (1976) demonstrates that organisations often escalate commitment to failing initiatives despite contradictory evidence. Such biases are particularly problematic during early-stage innovation because limited empirical evidence allows intuitive beliefs and organisational politics to dominate strategic decision-making. Consequently, evidence generation should not merely support innovation but actively challenge prevailing assumptions through structured experimentation and critical reflection.

Taken together, innovation management, organisational learning and behavioural decision-making literature converge on a common conclusion: successful digital innovation depends less upon predicting the future accurately than upon learning systematically under conditions of uncertainty. This perspective reframes product development as a cumulative process of organisational learning in which evidence, rather than implementation, constitutes the principal output of early-stage innovation. The subsequent evolution of digital product development methodologies can therefore be understood as a progressive attempt to improve how organisations learn before they build.

2.2 The Evolution of Digital Product Development

The evolution of digital product development over the past five decades reflects a gradual shift from predictive planning towards adaptive learning. This transformation has been driven by increasing environmental uncertainty, accelerating technological change and growing recognition that customer needs cannot always be specified accurately before development begins. Contemporary approaches to product management therefore represent not isolated methodological innovations but successive responses to the limitations of earlier development paradigms.

Early software engineering was dominated by sequential planning models, most notably the Waterfall methodology proposed by Royce (1970). Waterfall assumed that software requirements could be comprehensively defined before implementation, allowing development to progress through discrete phases of analysis, design, implementation, testing and deployment. Such approaches proved effective in relatively stable environments where customer requirements remained predictable and technological change was limited. However, they became increasingly problematic as software products grew more complex and markets more dynamic.

Subsequent empirical research demonstrated that software projects rarely conform to linear planning assumptions. Requirements evolve throughout development as customers gain experience, technologies mature and competitive conditions change. Under these circumstances, extensive upfront planning frequently creates rigidity rather than reducing uncertainty, increasing the cost of adaptation when assumptions prove incorrect. Consequently, project success became increasingly dependent on organisational responsiveness rather than adherence to predetermined plans.

The publication of the Agile Manifesto (Beck et al., 2001) represented a fundamental shift in software engineering philosophy. Rather than emphasising comprehensive documentation and predictive planning, Agile prioritised iterative development, customer collaboration, frequent software delivery and responsiveness to change. Numerous empirical studies have since demonstrated that Agile practices improve software quality, team productivity, stakeholder engagement and project adaptability, particularly within environments characterised by uncertainty and evolving requirements (Dingsøyr et al., 2012; Dikert, Paasivaara & Lassenius, 2016).

Agile's principal contribution lies in recognising that knowledge emerges throughout development rather than preceding it. By shortening feedback cycles and encouraging continuous customer interaction, Agile enables teams to incorporate learning into implementation decisions. Nevertheless, Agile primarily addresses uncertainty associated with software execution rather than uncertainty concerning product desirability or strategic relevance. Sprint reviews, retrospectives and iterative planning improve how software is developed but generally assume that the underlying product opportunity has already been identified and validated.

Parallel developments in software operations further extended these capabilities. DevOps integrated software development with operational management, enabling continuous integration, automated testing, continuous deployment and infrastructure automation. Research indicates that organisations adopting DevOps practices achieve significantly higher deployment frequency, shorter lead times, improved reliability and faster recovery from operational failures (Forsgren, Humble & Kim, 2018). Together, Agile and DevOps have substantially reduced implementation uncertainty by increasing organisational responsiveness and engineering efficiency.

More recently, the emergence of product operating models has further expanded the role of cross-functional teams within digital organisations. Rather than organising work around projects with predetermined deliverables, product operating models emphasise long-lived teams responsible for continuously delivering customer value through iterative learning, experimentation and adaptation (Cagan, 2020; Denning, 2018). This shift reflects growing recognition that digital products require continuous evolution rather than discrete project completion.

Despite these advances, a critical limitation remains. Most software engineering methodologies primarily optimise execution after a product direction has already been chosen. They provide comparatively limited guidance regarding how organisations should identify opportunities, formulate product strategies or evaluate competing investment alternatives before implementation begins. As Cooper (2019) observes, many new product failures originate long before engineering activities commence, arising instead from weak opportunity assessment, inadequate customer understanding or poor strategic alignment.

This limitation has become increasingly significant as advances in cloud computing, low-code platforms and generative AI have dramatically reduced the cost and speed of software production. Large language models now automate substantial components of software development, including code generation, documentation, testing and interface design (Noy & Zhang, 2023; Brynjolfsson, Li & Raymond, 2025). These developments have fundamentally altered the economics of software engineering by shifting the principal organisational constraint away from implementation capacity.

Paradoxically, the easier it becomes to build software, the more important it becomes to determine whether software should be built at all. Lower implementation costs reduce the penalty associated with experimentation but simultaneously increase the number of potential opportunities competing for organisational attention and investment. As Simon (1971) argued, abundance transforms the nature of scarcity. In an environment where software production becomes increasingly abundant, customer insight, strategic judgement and evidence-based prioritisation emerge as comparatively scarce organisational resources.

Consequently, contemporary digital innovation requires capabilities extending beyond efficient software engineering. Organisations must not only build products rapidly but also develop systematic mechanisms for identifying meaningful customer problems, evaluating market opportunities and generating evidence before significant development resources are committed. These emerging requirements have stimulated the evolution of product discovery as a distinct discipline within product management, which forms the focus of the following section.

2.3 From Product Delivery to Product Discovery

Although Agile and DevOps transformed software engineering by improving the efficiency and adaptability of product delivery, they provide comparatively limited guidance regarding how organisations should determine what products ought to be developed. This distinction has given rise to product discovery as a complementary discipline concerned with reducing uncertainty before substantial implementation commitments are made. Whereas product delivery seeks to build solutions efficiently, product discovery seeks to determine whether those solutions address meaningful customer problems, create strategic value and justify continued investment (Torres, 2021).

The emergence of product discovery reflects a broader shift within innovation management from execution-oriented thinking towards evidence-based learning. Historically, many organisations viewed product management primarily as a planning and coordination function responsible for requirements gathering, roadmap development and stakeholder communication. Product managers frequently operated as intermediaries between business stakeholders and software engineering teams, translating predetermined business requirements into technical specifications. While effective within relatively stable markets, this model assumes that customer needs can be identified accurately before development begins and remain sufficiently stable throughout implementation.

Increasing environmental uncertainty has challenged these assumptions. Digital markets evolve rapidly, customer expectations change continuously, and competitors frequently redefine market boundaries through technological innovation. Under such conditions, static product planning becomes increasingly problematic because assumptions formulated during initial planning may no longer reflect market realities by the time implementation is completed. Product management therefore increasingly emphasises learning, experimentation and adaptation rather than deterministic planning (Perri, 2018).

One of the most influential contributions to this shift is Blank's Customer Development methodology (Blank, 2013; Blank & Dorf, 2020). Developed initially within entrepreneurial contexts, Customer Development challenged the traditional assumption that product ideas should proceed directly into engineering following business planning. Instead, Blank argued that organisations should systematically test assumptions regarding customer problems, market segments, value propositions and business models before scaling development efforts. Customer interviews, hypothesis testing and iterative validation therefore became central mechanisms through which organisations progressively reduced uncertainty.

Building upon these ideas, Ries (2011) introduced the Lean Startup methodology, which reframed product development as a sequence of experiments designed to validate or invalidate assumptions under conditions of uncertainty. Central to Lean Startup is the Build–Measure–Learn feedback loop, through which organisations develop minimum viable products (MVPs), gather customer feedback and revise hypotheses accordingly. Rather than pursuing comprehensive planning, Lean Startup encourages rapid experimentation, iterative learning and evidence-based decision-making.

The influence of Lean Startup extends well beyond entrepreneurial ventures. Established organisations have increasingly adopted experimental approaches to innovation in recognition that assumptions regarding customer demand frequently prove inaccurate when subjected to empirical investigation. Numerous studies demonstrate that iterative experimentation reduces innovation risk by enabling organisations to identify flawed assumptions before committing substantial development resources (Eisenmann, Ries & Dillard, 2012; Shepherd & Gruber, 2021). Consequently, experimentation has become an increasingly accepted mechanism for managing uncertainty rather than merely evaluating completed products.

However, Lean Startup has also attracted criticism. Several scholars argue that its emphasis on rapid experimentation may encourage incremental optimisation rather than more fundamental innovation (McDonald & Eisenhardt, 2020). MVPs may provide useful evidence regarding product usability while failing to capture broader strategic considerations such as competitive positioning, organisational capability or long-term market evolution. Furthermore, many implementations focus heavily on validating existing product concepts rather than questioning whether the original opportunity itself should be pursued.

Similarly, Customer Development provides valuable mechanisms for customer engagement but offers comparatively limited guidance regarding organisational governance or strategic integration. Customer interviews frequently generate rich qualitative insights; however, organisations often struggle to translate these insights into coherent investment decisions that align with broader strategic objectives. Consequently, discovery activities may remain isolated from executive decision-making despite producing valuable evidence.

These limitations highlight an important distinction between conducting discovery activities and managing discovery as an organisational capability. Many organisations perform customer interviews, usability testing and experimentation intermittently while continuing to govern investment decisions primarily through executive judgement, roadmap commitments or political priorities. Discovery therefore becomes an operational activity rather than a strategic capability embedded within organisational decision-making.

More recently, Torres (2021) has sought to address this challenge through the concept of Continuous Discovery. Rather than treating customer research as an early-stage project activity, Continuous Discovery advocates ongoing customer engagement throughout the product lifecycle. Weekly customer interviews, continuous hypothesis refinement and regular opportunity assessment enable organisations to incorporate learning into product decisions continuously rather than episodically. This approach reflects broader developments within Agile thinking by recognising that customer understanding should evolve alongside products rather than preceding them.

Continuous Discovery represents an important conceptual advance because it positions learning as an enduring organisational activity rather than a temporary project phase. Nevertheless, it remains primarily focused on customer interactions and opportunity identification. Comparatively less attention is devoted to integrating customer evidence with organisational capabilities, competitive strategy, governance processes and executive decision-making. Discovery therefore remains only partially connected to broader theories of organisational adaptation.

A further limitation shared across many discovery methodologies concerns performance measurement. Organisations frequently continue evaluating product teams using delivery-oriented metrics such as sprint velocity, roadmap completion, feature throughput and release frequency (Perri, 2018). Although these measures provide valuable operational information, they reveal relatively little regarding whether organisations are learning effectively or reducing strategic uncertainty. Reward systems therefore continue encouraging execution rather than evidence generation, reinforcing a culture in which software production becomes an end in itself rather than a mechanism for creating customer value.

From an organisational learning perspective, this represents a significant misalignment. If the objective of discovery is uncertainty reduction, then progress should be evaluated according to the quality of knowledge generated rather than the quantity of software delivered. Discovery success should therefore be measured through validated assumptions, improved customer understanding, reduction in critical uncertainties and strengthened strategic confidence rather than simply through implementation outputs.

Collectively, these observations suggest that existing discovery methodologies provide valuable tools for experimentation and customer engagement but remain insufficiently integrated with organisational governance, strategic capability and evidence-based decision-making. The challenge is therefore not the absence of effective discovery techniques but the absence of a coherent architecture through which evidence accumulates, informs decisions and progressively shapes organisational strategy.

2.4 Human-Centred Innovation: Design Thinking, Human-Centred Design and Jobs-to-be-Done

Alongside the emergence of Lean Startup and Customer Development, human-centred approaches to innovation have significantly influenced contemporary product management by emphasising empathy, customer understanding and problem framing. These approaches collectively argue that successful innovation depends less upon technological sophistication than upon developing deep insights into the experiences, motivations and unmet needs of those whom products are intended to serve.

Among the most influential of these approaches is Design Thinking. Popularised by Brown (2009) and subsequently developed through both academic and practitioner research, Design Thinking conceptualises innovation as an iterative process involving empathising with users, defining problems, generating ideas, prototyping solutions and testing alternatives. Rather than beginning with technical feasibility, Design Thinking encourages organisations to investigate customer experiences before considering possible solutions.

Academic research increasingly supports the effectiveness of Design Thinking within uncertain innovation contexts. Liedtka (2018), for example, argues that Design Thinking reduces innovation risk by encouraging iterative experimentation and collaborative sense-making, enabling organisations to challenge assumptions before committing substantial resources. Cross (2021) similarly describes design as a discipline concerned fundamentally with navigating uncertainty through iterative problem exploration rather than deterministic solution development.

Despite its considerable influence, Design Thinking has attracted several criticisms. First, scholars argue that it often lacks theoretical precision, functioning more as a broad philosophy than as a rigorously specified decision framework (Johansson-Sköldberg, Woodilla & Çetinkaya, 2013). Organisations may adopt Design Thinking workshops and brainstorming exercises without embedding customer evidence within broader governance processes or strategic decision-making. Consequently, empathy becomes an isolated activity rather than an organisational capability.

Second, Design Thinking frequently prioritises creativity and idea generation over systematic hypothesis validation. While divergent thinking supports innovation by encouraging exploration of multiple alternatives, organisations ultimately require mechanisms for evaluating competing ideas through empirical evidence rather than creative consensus alone. The transition from inspiration to validated strategic decision-making therefore remains comparatively underdeveloped.

Closely related to Design Thinking is Human-Centred Design (HCD), formalised internationally through ISO 9241-210 (ISO, 2019). HCD emphasises designing products around users' characteristics, needs and contexts of use throughout the development lifecycle. Unlike traditional engineering approaches that prioritise technical functionality, HCD recognises usability, accessibility and user experience as fundamental determinants of product success.

Human-Centred Design has become particularly influential within digital product development because software products increasingly compete through customer experience rather than technical functionality alone. Norman (2013) argues that successful products emerge when design reflects users' cognitive models rather than technological possibilities. Consequently, HCD encourages iterative observation, prototyping and evaluation throughout development.

Nevertheless, HCD shares limitations similar to those identified within Design Thinking. Although it substantially improves customer understanding, it provides comparatively limited guidance regarding strategic prioritisation, competitive positioning or organisational capability. Organisations may develop highly usable products that nonetheless fail commercially because they address problems lacking sufficient strategic significance or market demand.

A complementary perspective is provided by Jobs-to-be-Done (JTBD) theory (Christensen et al., 2016; Ulwick, 2016). Rather than focusing on customer demographics or product features, JTBD argues that customers "hire" products to accomplish particular functional, emotional and social objectives. Innovation therefore depends upon understanding the underlying progress customers seek rather than the solutions they currently use.

JTBD represents an important conceptual advance because it shifts attention from products towards desired outcomes. This perspective encourages organisations to investigate why customers behave as they do rather than simply observing what they do. Outcome-Driven Innovation (Ulwick, 2016) further operationalises this perspective by providing structured methods for identifying underserved customer outcomes and prioritising innovation opportunities accordingly.

However, JTBD primarily explains customer demand rather than organisational decision-making. It offers valuable insights into opportunity identification but provides less guidance regarding how organisations should combine customer evidence with strategic capabilities, technological constraints and governance considerations. Like Design Thinking and HCD, JTBD contributes an important piece of the innovation puzzle without fully integrating it into a comprehensive organisational learning system.

Across these human-centred approaches, a common pattern emerges. All recognise that successful innovation begins with understanding customers rather than building technology. Each contributes valuable methods for uncovering latent needs, challenging assumptions and improving product desirability. Yet each also tends to focus predominantly on one dimension of innovation—customer understanding—without explicitly integrating this knowledge into broader organisational processes governing investment, capability development and strategic adaptation.

This observation reinforces an increasingly important conclusion emerging from the literature: no single methodology adequately addresses the full complexity of contemporary digital innovation. Design Thinking contributes empathy and creativity; Human-Centred Design improves usability; Jobs-to-be-Done clarifies customer outcomes; Lean Startup enables experimentation; Customer Development validates assumptions; Continuous Discovery institutionalises ongoing learning. Yet organisations continue to struggle because these complementary approaches often remain fragmented rather than functioning as components of a coherent evidence-based decision architecture.

2.5 Organisational Learning, Dynamic Capabilities and Evidence-Based Product Management

The preceding sections demonstrate that contemporary product discovery methodologies have substantially improved organisations' ability to understand customers, test assumptions and experiment with potential solutions. However, they also reveal a common limitation: the majority of these approaches focus on improving individual discovery activities rather than explaining how organisations systematically transform evidence into sustained strategic capability. Addressing this limitation requires a broader theoretical perspective grounded in organisational learning and dynamic capabilities.

Organisational learning theory has long argued that sustainable competitive advantage depends not only upon acquiring knowledge but also upon an organisation's capacity to interpret, disseminate and apply that knowledge effectively (Argyris & Schön, 1978; Senge, 1990). Learning therefore extends beyond individual experience to encompass organisational processes through which assumptions are questioned, decisions are revised and collective understanding evolves over time. Within innovation contexts, this distinction is particularly important because uncertainty cannot be eliminated through planning alone; instead, organisations must develop mechanisms that continuously transform new evidence into improved strategic decisions.

March's (1991) distinction between exploration and exploitation provides an influential foundation for understanding this challenge. Exploration involves experimentation, discovery, variation and the pursuit of new knowledge, whereas exploitation emphasises efficiency, refinement and execution. Although both activities are essential, organisations frequently struggle to balance them effectively. Excessive emphasis on exploitation promotes operational excellence but may reduce an organisation's capacity to recognise changing customer needs or emerging market opportunities. Conversely, excessive exploration risks generating numerous ideas without translating learning into commercial value.

Within digital product management this tension frequently manifests as an imbalance between discovery and delivery. Many organisations have invested heavily in Agile software development, DevOps and engineering productivity while devoting comparatively fewer resources to institutionalising systematic discovery practices. Consequently, they become highly proficient at exploiting established delivery capabilities while underinvesting in exploratory activities that generate strategic knowledge. Product discovery therefore risks becoming episodic rather than continuous, reducing its influence on organisational decision-making.

The concept of the learning organisation offers a complementary perspective. Senge (1990) argues that organisations capable of sustained adaptation exhibit several characteristics, including systems thinking, shared vision, reflective practice and a willingness to challenge deeply held assumptions. Rather than treating errors as failures, learning organisations regard unexpected outcomes as opportunities for improving understanding. This perspective closely aligns with evidence-based product discovery, where invalidated hypotheses represent valuable knowledge rather than unsuccessful projects.

Psychological safety further reinforces this capability. Edmondson (2019) demonstrates that organisations learn more effectively when individuals feel able to question assumptions, report contradictory evidence and challenge prevailing decisions without fear of negative consequences. Innovation therefore depends not only upon formal discovery processes but also upon organisational cultures that encourage intellectual humility and constructive disagreement. Where political incentives discourage challenging existing roadmaps or executive assumptions, evidence generation may become subordinate to organisational consensus, undermining the effectiveness of discovery activities.

Knowledge creation theory similarly emphasises that competitive advantage emerges through an organisation's ability to convert dispersed information into actionable knowledge (Nonaka & Takeuchi, 1995). Customer interviews, usability testing, market research and experimentation generate valuable information; however, competitive advantage depends upon integrating these disparate sources into coherent strategic understanding. Knowledge therefore represents more than accumulated data; it is evidence interpreted within organisational context to inform future action.

These perspectives converge on a central proposition: product discovery should not be understood as a discrete project phase but as an organisational learning capability embedded within governance, leadership and strategic decision-making. Discovery activities generate value only when evidence systematically influences investment decisions, strategic priorities and organisational adaptation. Consequently, organisations require frameworks that explicitly connect customer learning with broader organisational capability.

Dynamic capabilities theory provides precisely this strategic perspective. Teece (2018) argues that sustained competitive advantage depends upon three interrelated organisational capabilities: sensing opportunities, seizing opportunities and transforming organisational resources in response to environmental change. Importantly, these activities are inseparable. Identifying attractive opportunities without the capability to exploit them creates limited value, while efficient execution without accurate sensing risks optimising products that customers do not require.

This insight has profound implications for digital product discovery. Existing methodologies largely emphasise sensing customer needs through interviews, experimentation and observation. Comparatively less attention has been devoted to integrating customer evidence with organisational capabilities, competitive positioning and strategic priorities. Consequently, organisations may identify attractive customer opportunities that remain commercially unattractive because they fail to align with distinctive organisational strengths.

The resource-based view further reinforces this argument. Barney (1991) contends that competitive advantage derives not from widely available technologies but from valuable, rare and difficult-to-imitate organisational capabilities. As cloud computing, open-source software and generative AI progressively democratise software development, technical implementation becomes increasingly commoditised. Competitive differentiation therefore shifts towards capabilities relating to customer understanding, organisational learning and evidence-based strategic decision-making. These capabilities are substantially more difficult for competitors to replicate because they depend upon organisational routines, culture and accumulated knowledge rather than technological assets alone.

Evidence-based management provides the final component of this theoretical foundation. Rousseau (2006) argues that managerial decisions should integrate scientific evidence, organisational data, practitioner expertise and stakeholder values rather than relying predominantly upon intuition or authority. Within product management, however, evidence-based decision-making remains comparatively underdeveloped. Organisations frequently gather extensive customer feedback while continuing to prioritise roadmap commitments, executive preferences or historical precedent when allocating resources. Discovery therefore produces information without necessarily improving decisions.

Collectively, organisational learning, dynamic capabilities and evidence-based management suggest that successful innovation depends upon more than conducting discovery activities. It requires organisational systems through which evidence accumulates, challenges assumptions and progressively reshapes strategic direction. This observation provides the theoretical foundation upon which the HDNA Framework is constructed.

2.6 Artificial Intelligence and the Changing Economics of Product Discovery

The rapid emergence of generative artificial intelligence represents one of the most significant developments in digital innovation since the introduction of Agile software development. Large language models, code-generation systems and AI-assisted design tools have substantially reduced the effort required to produce software artefacts, documentation, prototypes and analytical outputs. Consequently, AI has become an increasingly important component of software engineering practice.

Current research overwhelmingly emphasises AI's productivity effects. Experimental evidence demonstrates that generative AI improves performance across a range of knowledge-intensive tasks, including programming, writing and information synthesis (Noy & Zhang, 2023). Similarly, Brynjolfsson, Li and Raymond (2025) report substantial productivity gains among professionals adopting AI-assisted workflows, particularly where tasks involve information processing and routine cognitive activities. These findings suggest that AI reduces implementation costs while increasing organisational capacity to produce software rapidly.

Although these operational benefits are substantial, they represent only part of AI's strategic significance. Most existing literature conceptualises AI as improving execution rather than altering the nature of innovation itself. This paper argues that AI fundamentally changes the distribution of organisational value creation by shifting the primary constraint within digital product development.

Historically, software engineering represented the principal bottleneck in digital innovation. Building sophisticated software required specialised technical expertise, lengthy development cycles and considerable financial investment. Consequently, organisations competed largely through engineering capability and implementation efficiency. Advances in Agile methodologies, cloud infrastructure and DevOps progressively reduced these constraints, while generative AI has accelerated this transformation by automating significant components of software production.

As implementation becomes increasingly abundant, scarcity shifts elsewhere. Simon (1971) argued that information abundance creates scarcity of attention; similarly, AI-generated software abundance creates scarcity of strategic judgement. Organisations can now generate far more ideas, prototypes and software solutions than they can meaningfully evaluate. The limiting factor therefore becomes not production capacity but the ability to determine which opportunities deserve investment.

This shift fundamentally elevates the importance of product discovery. AI can generate product concepts, summarise customer interviews, analyse qualitative feedback, create prototypes and recommend technical architectures. However, it cannot independently establish whether a customer problem is sufficiently important, whether an opportunity aligns with organisational capabilities or whether long-term strategic investment is justified. These questions require contextual understanding, organisational judgement and empirical validation through interaction with real customers.

Consequently, AI should be understood primarily as an augmentation technology rather than a replacement for discovery. Its greatest contribution may lie not in automating strategic decisions but in enabling organisations to generate and evaluate evidence more rapidly. AI can synthesise large volumes of customer data, identify emerging behavioural patterns, support competitive intelligence and accelerate experimentation, thereby expanding organisational learning capacity without replacing human interpretation.

Paradoxically, therefore, AI increases rather than decreases the strategic importance of product discovery. By lowering implementation costs, AI reduces the penalty associated with building software while simultaneously increasing the cost of selecting inappropriate opportunities. Organisations capable of learning faster than competitors are therefore likely to derive greater value from AI than organisations that merely automate software production.

This perspective extends current discussions regarding AI beyond productivity enhancement towards broader theories of organisational adaptation. AI changes not only how software is built but also where sustainable competitive advantage resides.

2.7 Critical Synthesis and Research Gap

The literature reviewed throughout this chapter demonstrates remarkable progress in understanding digital innovation. Agile software development has transformed implementation by improving adaptability and reducing execution uncertainty. Lean Startup and Customer Development have introduced systematic experimentation and hypothesis validation into entrepreneurial practice. Design Thinking, Human-Centred Design and Jobs-to-be-Done have deepened understanding of customer needs, while organisational learning and dynamic capabilities have explained how organisations adapt under conditions of uncertainty. More recently, generative AI has dramatically reduced the cost and speed of software development.

Despite these advances, the literature remains fragmented.

Each research tradition addresses a particular dimension of innovation uncertainty while providing comparatively limited guidance regarding how these dimensions should be integrated within organisational decision-making. Agile improves delivery but assumes that product direction has already been established. Lean Startup validates business assumptions but provides relatively little guidance concerning strategic capability. Design Thinking enhances empathy yet offers limited mechanisms for governance. Human-Centred Design improves usability but does not explicitly address organisational strategy. Jobs-to-be-Done explains customer demand but not organisational adaptation. Organisational learning explains knowledge creation but does not prescribe structured discovery processes. Evidence-based management advocates empirical decision-making but remains largely disconnected from contemporary product management practice.

Consequently, organisations frequently adopt multiple complementary methodologies simultaneously without possessing an overarching architecture that explains how evidence should accumulate across discovery activities or inform strategic investment decisions. Customer interviews, usability testing, experimentation and roadmap planning often occur independently, reducing the cumulative value of organisational learning. Discovery therefore becomes a collection of techniques rather than an integrated organisational capability.

The emergence of generative AI further exposes this theoretical limitation. Existing innovation frameworks were largely developed when software implementation remained comparatively expensive. As AI progressively automates development activities, the relative importance of customer understanding, strategic judgement and uncertainty reduction increases. Yet current literature has only begun to examine how this technological transformation reshapes the relationship between discovery and delivery.

Accordingly, this review identifies two interrelated gaps.

The first is a theoretical integration gap. While numerous methodologies contribute valuable approaches to product discovery, no comprehensive conceptual framework currently integrates strategic context, customer understanding, hypothesis generation, evidence evaluation and organisational learning within a coherent evidence-governed decision architecture.

The second is an AI-enabled strategic gap. Existing research has primarily examined AI as a productivity-enhancing technology, with comparatively limited attention devoted to its implications for product discovery, organisational learning and strategic decision-making. As software generation becomes increasingly commoditised, understanding how organisations should generate, interpret and govern evidence becomes an increasingly important research challenge.

To address these gaps, the following chapter introduces the Hypothesis-Driven Needs Assessment (HDNA) Framework. Rather than proposing another standalone innovation methodology, HDNA synthesises complementary theories from Agile development, Lean Startup, Customer Development, Human-Centred Design, Continuous Discovery, organisational learning, dynamic capabilities and evidence-based management into a unified framework for reducing uncertainty throughout digital product discovery. In doing so, it positions evidence—not software—as the principal output of early-stage innovation and organisational learning as the fundamental mechanism through which sustainable digital innovation is achieved.

3. The HDNA Framework: An Evidence-Based Model for Digital Product Discovery

3.1 Conceptual Foundation

The literature reviewed in Chapter Two demonstrates that contemporary product discovery is characterised by methodological richness but conceptual fragmentation. Existing approaches—including Agile software development, Lean Startup, Customer Development, Design Thinking, Human-Centred Design, Jobs-to-be-Done, Continuous Discovery and evidence-based management—each make important contributions to reducing uncertainty during innovation. However, these approaches typically address discrete aspects of the discovery process rather than providing a unified framework that explains how evidence should be systematically generated, integrated and used to guide strategic decision-making throughout the lifecycle of digital product development.

This fragmentation presents a significant challenge for organisations operating in increasingly uncertain and technology-intensive environments. In practice, organisations rarely rely on a single methodology. Instead, they combine customer interviews, design workshops, hypothesis testing, experimentation, roadmap planning and Agile delivery in ways that often lack an overarching conceptual logic. As a result, valuable insights generated through discovery activities frequently remain disconnected from strategic governance, investment decisions and organisational learning. Discovery becomes a collection of isolated techniques rather than a coherent organisational capability for reducing uncertainty.

The HDNA (Hypothesis-Driven Needs Assessment) Framework is proposed as a conceptual response to this limitation. Rather than introducing another standalone innovation methodology, HDNA provides an integrative decision architecture that organises established discovery practices into a coherent evidence-generation process. Its objective is not to replace existing approaches but to explain how they can operate collectively as complementary mechanisms for transforming uncertainty into progressively stronger organisational knowledge. The framework therefore functions as a meta-framework, synthesising established theories into a unified model for evidence-based product discovery.

At its core, HDNA is founded upon the proposition that digital product development should be understood primarily as a process of organisational learning rather than software production. Although software remains the visible outcome of innovation, the more fundamental objective of product discovery is to improve the quality of strategic decisions before substantial implementation resources are committed. In this respect, the framework extends the evidence-based management perspective advanced by Rousseau (2006) into the domain of digital product innovation by treating every significant product decision as an empirical proposition requiring systematic investigation rather than managerial consensus.

This perspective reflects the growing recognition that uncertainty is the defining characteristic of innovation. Unlike routine operational activities, where objectives and procedures are largely known in advance, innovation requires organisations to make decisions under conditions of incomplete information, ambiguous customer needs and continuously evolving technological and competitive environments (March, 1991; Teece, 2018). Under such circumstances, planning alone cannot eliminate uncertainty. Instead, uncertainty must be progressively reduced through iterative cycles of observation, experimentation, reflection and evidence generation.

The HDNA Framework is therefore underpinned by three interrelated theoretical principles.

First, innovation is fundamentally a process of uncertainty reduction. Product teams rarely begin with validated knowledge regarding customer needs, market opportunities or solution effectiveness. Instead, they begin with assumptions. The purpose of discovery is to progressively transform those assumptions into evidence, enabling organisations to increase confidence in strategic decisions before committing substantial investment. Progress is therefore measured not by the volume of features produced, but by the quality of knowledge generated.

Second, organisational learning is cumulative and recursive. Knowledge generated during discovery should not be treated as isolated project outputs but as organisational assets that continually reshape strategic understanding. This principle draws directly upon Argyris and Schön's (1978) concept of double-loop learning, in which organisations question not only operational practices but also the underlying assumptions that inform strategic decisions. Similarly, it reflects Teece's (2018) dynamic capabilities framework, in which organisations continuously sense emerging opportunities, seize promising directions and transform their capabilities in response to new evidence. Consequently, every discovery activity within HDNA contributes to an evolving organisational understanding rather than merely validating individual product ideas.

Third, effective governance depends upon making uncertainty explicit rather than assuming certainty prematurely. Conventional product governance frequently relies upon business cases, roadmaps and implementation milestones that imply a level of confidence unsupported by empirical evidence. HDNA instead requires assumptions to be explicitly articulated, prioritised and systematically tested before investment decisions are made. By exposing uncertainty rather than concealing it, organisations create conditions for more rational decision-making and reduce the likelihood of confirmation bias, escalation of commitment and politically motivated investment decisions.

Collectively, these principles reposition product discovery from a preliminary design activity to a strategic organisational capability. Within the HDNA Framework, discovery is not simply the phase that precedes delivery; rather, it is the mechanism through which organisations continuously generate, evaluate and integrate evidence to improve strategic judgement. In doing so, the framework reframes the primary output of early-stage innovation. Instead of software, prototypes or product specifications, the principal output becomes validated organisational knowledge capable of informing progressively more confident investment decisions.

3.2 Overview of the HDNA Framework

Building upon these conceptual foundations, the HDNA Framework structures product discovery as an iterative sequence of four interconnected phases: Envisioning, Strategy, Ideation and Detailing. Each phase addresses a distinct category of uncertainty while contributing to a cumulative evidence base that supports progressively stronger strategic decision-making. Rather than representing a linear project methodology, the framework should be understood as a dynamic learning system in which evidence continuously informs subsequent decisions and may require earlier assumptions to be revisited as new insights emerge.

The sequencing of the four phases reflects the progressive refinement of organisational knowledge. Innovation begins with a broad exploration of context before progressively narrowing towards increasingly specific product decisions. This progression recognises that uncertainty cannot be eliminated in a single step; instead, different forms of uncertainty must be addressed at different stages of discovery. Strategic uncertainty regarding markets and organisational capabilities precedes uncertainty concerning customer problems; customer uncertainty precedes solution uncertainty; and solution uncertainty precedes implementation uncertainty. Attempting to resolve these questions out of sequence frequently results in premature commitment to poorly understood opportunities.

The first phase, Envisioning, establishes the strategic context within which innovation occurs. Rather than generating solutions immediately, organisations seek to understand the broader environment by examining customer ecosystems, technological developments, competitive dynamics, regulatory influences and organisational capabilities. The objective is to develop an informed understanding of where meaningful opportunities may exist before formulating specific product concepts.

The second phase, Strategy, transforms this contextual understanding into explicit product hypotheses. Here, organisations formulate and prioritise assumptions regarding target customers, customer problems, value propositions and strategic differentiation. Importantly, these assumptions are treated not as strategic conclusions but as hypotheses requiring empirical validation. Strategy therefore becomes an evidence-informed process of opportunity formulation rather than traditional business planning.

The third phase, Ideation, shifts attention from problem definition towards potential solutions. Guided by the strategic hypotheses established during the preceding phase, teams generate multiple solution alternatives, rapidly prototype promising concepts and evaluate them through structured experimentation with representative users. Rather than seeking immediate convergence on a preferred design, the objective is to identify which solution concepts most effectively address validated customer needs while remaining consistent with organisational strategy.

The fourth phase, Detailing, translates validated concepts into implementation-ready artefacts without abandoning the principles of evidence generation established earlier in the framework. Technical architecture, interaction design, product requirements and delivery planning continue to evolve through usability testing, feasibility assessment and iterative refinement. Discovery and delivery therefore remain closely coupled, ensuring that implementation decisions continue to be informed by empirical learning rather than static specifications.

Although these four phases provide a useful conceptual structure, the defining characteristic of HDNA lies not in the phases themselves but in the learning loops that connect them. Existing innovation methodologies include many comparable activities; however, they frequently lack explicit mechanisms through which evidence accumulates across discovery activities and informs organisational governance. HDNA addresses this limitation by embedding structured evidence reviews between successive phases. Progression is determined not by task completion but by the strength of evidence supporting continued investment.

Consequently, the framework transforms traditional stage-gate governance into an evidence-governed learning process. Instead of asking whether predetermined deliverables have been completed, decision-makers ask whether sufficient knowledge has been generated to justify advancing to the next stage of investment. This seemingly subtle distinction fundamentally changes how innovation is governed. Progress becomes synonymous with increasing decision confidence rather than increasing implementation effort.

The cumulative effect is an organisational system that continuously converts assumptions into evidence and evidence into strategic knowledge. By explicitly integrating customer understanding, strategic reasoning, experimentation and organisational learning, HDNA provides a coherent architecture for evidence-based digital product discovery that complements existing delivery methodologies while addressing the strategic uncertainties they largely assume to have been resolved.

3.3 Phase One: Envisioning

The HDNA Framework deliberately begins not with ideas or solutions, but with context. This reflects a central argument developed throughout the preceding chapters: organisations frequently fail not because they execute poor solutions, but because they commit prematurely to solving poorly understood problems. Effective product discovery therefore requires organisations to first develop a comprehensive understanding of the environment within which innovation occurs before attempting to define specific product opportunities.

The purpose of the Envisioning phase is to establish this strategic context by systematically exploring the external and internal factors that shape future innovation opportunities. Rather than evaluating individual product concepts, organisations seek to understand broader patterns of technological change, customer behaviour, market evolution, competitive dynamics and organisational capability. In doing so, Envisioning functions as the primary sensing mechanism within the HDNA Framework, providing the knowledge base from which subsequent discovery activities emerge.

This orientation is strongly informed by Dynamic Capabilities theory (Teece, 2018), which argues that sustained competitive advantage depends upon an organisation's ability to sense emerging opportunities before competitors. Sensing involves more than environmental scanning; it requires interpreting weak signals, identifying changing customer expectations and recognising how technological developments may alter existing markets. Within digital innovation, where product lifecycles are increasingly compressed and technological disruption is continuous, this capability becomes particularly significant. Organisations that fail to invest in systematic sensing risk optimising products for markets that no longer exist or overlooking opportunities created by emerging technologies such as generative AI.

Envisioning also draws upon strategic foresight and organisational learning literature by recognising that innovation begins with interpretation rather than prediction. The objective is not to forecast the future with certainty but to develop sufficiently rich contextual understanding to identify plausible opportunity spaces worthy of further investigation. This distinction is important because innovation rarely emerges from isolated customer observations or technological trends alone. Rather, opportunities arise through the interaction of customer needs, technological possibilities, competitive conditions and organisational strengths. Consequently, Envisioning encourages organisations to synthesise diverse sources of information into an integrated understanding of the innovation landscape.

Activities undertaken during this phase typically include environmental scanning, market and industry analysis, customer ecosystem mapping, technology assessment, competitor analysis, regulatory review and organisational capability evaluation. Unlike conventional strategic planning exercises, however, these activities are not intended to produce definitive strategic conclusions. Instead, they generate explicit assumptions regarding where value may be created, which customer groups appear underserved, what organisational capabilities may provide competitive advantage and which uncertainties warrant further investigation.

The principal outputs of Envisioning therefore include opportunity domains, contextual maps, capability assessments, emerging customer hypotheses and strategic assumptions. Importantly, none of these outputs are regarded as validated knowledge. They represent the organisation's current understanding of the innovation landscape and serve as the starting point for subsequent hypothesis development rather than the endpoint of strategic analysis. By explicitly distinguishing assumptions from evidence, the framework reinforces the principle that strategic confidence must be earned through empirical investigation rather than inferred from analytical sophistication alone.

This emphasis on context provides an important point of differentiation from many existing product discovery methodologies, which often begin with customer problems or solution ideation. While customer understanding remains essential, HDNA argues that customer needs cannot be fully interpreted without reference to broader organisational and environmental conditions. Opportunities that appear attractive from a customer perspective may ultimately prove strategically unviable if they fail to align with organisational capabilities or emerging market dynamics. Conversely, technological or competitive developments may reveal opportunities that are not immediately apparent through customer research alone. Envisioning therefore broadens the scope of discovery by positioning customer insight within a wider strategic context.

Finally, the Envisioning phase establishes the epistemological foundation upon which the remainder of the framework is constructed. By making assumptions explicit at the outset, subsequent phases are able to evaluate, refine or reject them through systematic evidence generation. Rather than beginning with confidence, the framework begins with informed uncertainty. This shift in perspective is fundamental to HDNA's conception of evidence-based product discovery and underpins every subsequent stage of the framework.

3.4 Phase Two: Strategy

Following the establishment of strategic context, the second phase of the HDNA Framework focuses on determining whether identified opportunities warrant further investigation. Whereas the Envisioning phase seeks to understand the external and internal environment within which innovation occurs, the Strategy phase transforms that contextual understanding into explicit, testable propositions regarding where organisational value may be created.

Unlike conventional strategic planning, this phase does not attempt to produce a comprehensive business case or definitive product roadmap. Such artefacts often create an illusion of certainty at precisely the point where uncertainty remains greatest. Instead, HDNA conceptualises strategy as a process of disciplined hypothesis formulation. Product strategy therefore becomes an evolving explanation of why a particular opportunity is expected to create customer and organisational value rather than a fixed commitment to a predetermined course of action.

This perspective reflects contemporary research on product strategy, which increasingly argues that successful product organisations distinguish strategic intent from implementation decisions (Cagan, 2020; Perri, 2018). Traditional planning approaches frequently assume that customer needs, market conditions and competitive dynamics can be sufficiently understood before development begins. However, as demonstrated throughout Chapter Two, digital markets are characterised by continual change, making static planning progressively less reliable. Under such conditions, product strategy must remain adaptive, evolving in response to new evidence rather than becoming constrained by historical assumptions.

Within HDNA, strategy is therefore expressed through an integrated product hypothesis comprising four interdependent components:

• the target customer whose needs the organisation intends to address;

• the customer problem or unmet need requiring investigation;

• the proposed value proposition explaining how meaningful value may be created; and

• the strategic differentiation through which the organisation expects to create sustainable advantage.

Collectively, these elements establish the conceptual foundation upon which subsequent discovery activities are organised. Importantly, none of these propositions are regarded as validated conclusions. Instead, each represents an explicit assumption whose validity must be demonstrated through empirical investigation.

This distinction fundamentally changes the role of strategy within product discovery. Rather than producing certainty, strategy provides structure for learning. By articulating assumptions explicitly, organisations create a shared understanding of precisely what requires validation before significant investment decisions are made. Strategy therefore functions not as the endpoint of planning but as the starting point for systematic organisational learning.

To support this process, organisations employ a range of complementary discovery activities. Exploratory customer interviews, stakeholder workshops, competitor analysis, market mapping and opportunity assessment contribute different forms of evidence that collectively strengthen—or weaken—the evolving product hypothesis. Consistent with Blank's (2013) Customer Development methodology and Ries' (2011) Lean Startup philosophy, evidence is accumulated iteratively rather than through a single validation event. As understanding improves, hypotheses are refined, reformulated or discarded according to the quality of empirical evidence rather than organisational preference.

Importantly, HDNA extends existing discovery methodologies by integrating strategic capability into the evaluation process. Customer desirability alone is insufficient to justify continued investment. Attractive customer opportunities may remain strategically inappropriate if they fail to align with organisational resources, technological capabilities or long-term competitive positioning. Conversely, opportunities that leverage distinctive organisational strengths may justify continued exploration despite initially ambiguous customer evidence. Product strategy therefore emerges from the interaction between customer understanding and organisational capability rather than either dimension in isolation.

The principal outputs of the Strategy phase include a refined opportunity statement, prioritised assumptions, an articulated product hypothesis, an initial value proposition and a structured evidence plan identifying which uncertainties require investigation during subsequent discovery. These outputs provide the intellectual foundation for the Ideation phase, ensuring that solution development remains anchored to validated strategic reasoning rather than creative intuition alone.

3.5 Phase Three: Ideation

Having established a coherent strategic hypothesis, the HDNA Framework progresses towards the exploration of potential solutions. Unlike many innovation methodologies, however, Ideation is not conceived primarily as a creative exercise. Rather, it represents the systematic generation of experimental artefacts through which strategic assumptions regarding customer value can be investigated.

This distinction is fundamental. Traditional innovation processes frequently position ideation as the generation of promising ideas, implicitly encouraging teams to identify the "best" solution as early as possible. Such an approach risks premature convergence, whereby organisations become increasingly committed to attractive concepts before obtaining sufficient evidence that those concepts genuinely address meaningful customer needs. HDNA instead treats solution concepts as competing hypotheses regarding how validated customer problems might be resolved.

This perspective draws together several complementary theoretical traditions reviewed in Chapter Two. Design Thinking contributes techniques for divergent exploration and human-centred creativity (Brown, 2009); Human-Centred Design provides structured principles for iterative user evaluation (ISO 9241-210; Norman, 2013); Continuous Discovery institutionalises ongoing customer engagement throughout product development (Torres, 2021); while Lean Startup emphasises experimentation as the primary mechanism for reducing uncertainty (Ries, 2011). HDNA synthesises these approaches by positioning every prototype, concept or experiment as an instrument for generating evidence rather than demonstrating design capability.

Consequently, ideation becomes an exercise in comparative learning. Multiple alternative solutions are intentionally developed because competing concepts reveal different assumptions about customer behaviour, value creation and usability. Rather than asking which design appears most innovative, organisations ask which explanation of customer behaviour is most strongly supported by empirical evidence.

Experimental methods employed during this phase may include low-fidelity prototypes, interactive wireframes, concept visualisations, Wizard-of-Oz experiments, concierge MVPs, landing-page experiments, usability testing, behavioural observation and simulated service interactions. Although these techniques differ considerably in complexity, they share a common purpose: generating credible evidence regarding customer desirability while minimising unnecessary implementation effort.

Importantly, HDNA deliberately discourages emotional attachment to individual solutions. Within many organisations, product ideas rapidly become associated with particular teams or senior stakeholders, creating psychological barriers to abandoning weak concepts. By explicitly framing prototypes as temporary experiments rather than emerging products, the framework reduces escalation of commitment and encourages intellectual flexibility. Invalidated concepts are therefore interpreted not as failures but as valuable contributions to organisational knowledge.

The evidence generated throughout Ideation extends beyond traditional usability assessment. Customer interactions reveal assumptions regarding willingness to adopt, perceived value, behavioural change, decision-making processes and contextual influences that frequently remain invisible within conventional market research. Consequently, experimentation contributes not only to improving individual solutions but also to refining the strategic hypotheses established during the preceding phase.

An important feature of HDNA is the recursive relationship between Strategy and Ideation. Experimental evidence may strengthen confidence in the original product hypothesis, require modifications to specific assumptions or fundamentally challenge the opportunity itself. Learning therefore flows in both directions. Rather than progressing through a rigid sequence of predefined stages, organisations move iteratively between strategic reasoning and empirical observation until sufficient evidence has accumulated to justify increased investment.

By the conclusion of the Ideation phase, uncertainty surrounding customer desirability should be substantially reduced. Organisations possess not merely a preferred solution concept but, more importantly, a stronger evidential basis for explaining why that solution is expected to create customer value. The primary outcome is therefore increased strategic confidence rather than increased design sophistication.

3.6 Phase Four: Detailing

Only after substantial uncertainty regarding strategic direction and customer desirability has been addressed does the HDNA Framework transition towards implementation planning. This sequencing reflects one of the framework's central propositions: implementation should be the consequence of learning rather than the mechanism through which learning occurs.

Conventional product development methodologies frequently treat specification as the point at which discovery concludes and execution begins. Once requirements have been documented, uncertainty is implicitly assumed to have been resolved, and organisational attention shifts towards efficient delivery. HDNA challenges this assumption. Although uncertainty has been substantially reduced by this stage, it has not disappeared. Technical feasibility, implementation complexity, usability, scalability and organisational readiness remain active sources of uncertainty requiring continued investigation.

Accordingly, the Detailing phase does not represent a transition from discovery to delivery but rather an evolution in the nature of evidence generation. Whereas earlier phases primarily investigate strategic and customer uncertainty, Detailing concentrates on implementation uncertainty. The objective is to translate validated concepts into development-ready artefacts while maintaining the evidence-based principles established throughout the framework.

Typical outputs include interaction models, information architecture, service blueprints, technical architecture, product requirements, development backlogs, implementation roadmaps and design system specifications. Importantly, these artefacts are regarded as evolving representations of current organisational knowledge rather than immutable project documentation. As new technical or customer evidence emerges, they remain open to revision.

Evidence generation therefore continues throughout Detailing. Technical spikes explore architectural feasibility, usability testing evaluates increasingly realistic prototypes, accessibility reviews ensure inclusive design, and implementation experiments identify engineering risks before full-scale development commences. This approach closely aligns with contemporary product engineering practices in which discovery and delivery increasingly occur concurrently rather than as sequential organisational functions.

Maintaining discovery activities throughout implementation planning provides several important organisational benefits. First, it reduces the likelihood that technical constraints remain hidden until development is well underway, thereby avoiding costly redesign. Second, it preserves close alignment between engineering decisions and validated customer needs, ensuring that implementation choices continue to support the strategic intent established during earlier discovery phases. Finally, it reinforces the principle that knowledge creation remains an ongoing organisational capability rather than an activity confined to the beginning of projects.

The completion of the Detailing phase therefore signifies more than the production of implementation documentation. It represents the point at which the organisation has accumulated sufficient strategic, customer and technical evidence to justify committing substantial development resources with an informed level of confidence. Rather than eliminating uncertainty entirely, the framework ensures that the remaining uncertainties are understood, explicitly acknowledged and proportionate to the scale of the proposed investment.

In this respect, Detailing completes the transition from opportunity exploration to implementation readiness while preserving the evidence-based philosophy that distinguishes HDNA from more conventional product development approaches. Implementation is no longer viewed as the stage at which assumptions become fixed; instead, it becomes the continuation of an organisational learning process that extends throughout the entire product lifecycle.

3.7 Learning Loops: The Integrative Mechanism of the HDNA Framework

Although the HDNA Framework is organised into four sequential phases, its principal theoretical contribution does not reside in the phases themselves. Comparable activities—including environmental scanning, strategic planning, ideation and product specification—are widely represented across existing innovation methodologies. What distinguishes HDNA is the explicit integration of these activities through structured learning loops that systematically govern the accumulation and evaluation of evidence throughout the discovery process.

The learning loop represents the central coordinating mechanism of the framework. Rather than treating phase transitions as administrative milestones or project approvals, each transition becomes an opportunity for organisational reflection in which previously generated evidence is critically evaluated before additional resources are committed. This shifts governance away from monitoring project progress towards evaluating the quality of organisational knowledge.

This distinction is significant because traditional product governance frequently measures progress using implementation-oriented indicators such as feature completion, sprint velocity, roadmap adherence or delivery milestones. While these measures provide valuable information regarding execution efficiency, they reveal comparatively little about whether an organisation has improved its understanding of customer needs, reduced strategic uncertainty or increased confidence in future investment decisions. Consequently, organisations may appear operationally successful while continuing to pursue fundamentally flawed opportunities.

HDNA addresses this limitation by redefining progress as the progressive reduction of uncertainty. At the conclusion of each phase, teams are required to synthesise accumulated evidence and critically evaluate its implications for subsequent decision-making. Four guiding questions structure every learning review:

1. Which assumptions have been sufficiently supported by available evidence?

2. Which critical uncertainties remain unresolved?

3. What evidence justifies our current strategic direction?

4. Does the available evidence warrant additional organisational investment?

Collectively, these questions operationalise the principles of evidence-based management (Rousseau, 2006) within the context of digital product discovery. Rather than allowing organisational momentum or executive preference to determine project continuation, progression becomes contingent upon the quality, consistency and credibility of the evidence accumulated throughout discovery.

Importantly, the learning loop is designed to facilitate organisational learning rather than merely project control. Consistent with the theories of Argyris and Schön (1978), evidence may not only validate operational assumptions but also challenge the strategic beliefs that originally motivated the initiative. Customer research may reveal that the targeted problem lacks sufficient significance; competitive analysis may undermine assumptions regarding market differentiation; technological investigations may expose implementation constraints that fundamentally alter the attractiveness of an opportunity. By explicitly encouraging such reflection, the framework promotes double-loop learning, enabling organisations to revise underlying strategic assumptions rather than simply improving execution within an existing strategy.

The learning loops also reinforce the dynamic capabilities perspective introduced in Chapter Two. Teece (2018) argues that organisations sustain competitive advantage through their capacity to sense emerging opportunities, seize attractive possibilities and continuously transform organisational capabilities in response to environmental change. HDNA operationalises these capabilities through an iterative governance process in which sensing activities generate evidence, strategic evaluation determines whether opportunities should be seized, and learning informs ongoing organisational transformation. Consequently, discovery becomes embedded within organisational adaptation rather than remaining an isolated project activity.

A defining feature of the learning loop is its recognition that evidence may legitimately support multiple strategic outcomes. Following each review, three alternative decisions are possible:

Proceed. Available evidence provides sufficient confidence to justify progressing to the subsequent phase and committing additional organisational resources.

Iterate. Existing evidence remains incomplete or internally inconsistent, indicating that further investigation is required before responsible progression can occur.

Pivot or Terminate. Newly generated evidence contradicts critical assumptions underlying the current opportunity, indicating that resources should be redirected towards a revised strategic direction or withdrawn entirely.

The explicit inclusion of termination as a positive learning outcome represents an important departure from many conventional innovation processes. Within traditional governance models, project continuation is frequently interpreted as evidence of organisational success, while termination is perceived as failure. Such incentives contribute to escalation of commitment, whereby organisations continue investing in weak initiatives because previous investments create psychological, political or financial barriers to change (Staw, 1976). HDNA instead recognises that abandoning an unsupported hypothesis constitutes a successful outcome when it prevents significantly larger investments in products that lack sufficient strategic or customer value.

Accordingly, the learning loop transforms governance from a mechanism of project oversight into a mechanism of organisational learning. Evidence does not merely document progress; it determines progress. This distinction represents the conceptual core of the HDNA Framework and differentiates it from methodologies that primarily optimise discovery activities without explicitly governing how discovery informs strategic investment decisions.

3.8 Theoretical Propositions

As a conceptual framework, HDNA is intended not only to organise existing knowledge but also to stimulate future empirical investigation. The framework therefore gives rise to several theoretically derived propositions concerning the relationship between evidence generation, organisational learning and digital innovation performance. These propositions provide an initial research agenda through which the framework may be refined, tested and extended across different organisational contexts.

Proposition 1 (P1)

Organisations that systematically integrate evidence generation throughout product discovery will achieve superior product-market fit than organisations whose innovation processes remain primarily feature or delivery driven.

This proposition follows directly from the framework's central assumption that reducing uncertainty before implementation improves the alignment between organisational investment and customer value creation.

Proposition 2 (P2)

The explicit identification, documentation and continual revision of strategic assumptions will improve organisational decision quality by reducing confirmation bias and increasing reflective learning during product discovery.

By making assumptions visible, organisations become better able to evaluate contradictory evidence and avoid reinforcing unsupported strategic beliefs.

Proposition 3 (P3)

Evidence-governed learning loops will reduce unnecessary development investment by identifying weak product opportunities earlier in the innovation process than conventional stage-gate governance.

Rather than evaluating projects according to documentation or milestone completion, organisations allocate resources according to progressively increasing evidential confidence.

Proposition 4 (P4)

As generative artificial intelligence reduces the cost and effort of software implementation, the relative strategic importance of product discovery capabilities will increase.

This proposition reflects the central argument advanced throughout this dissertation: technological abundance shifts organisational scarcity away from implementation capacity towards customer understanding, strategic judgement and evidence-based decision-making.

Proposition 5 (P5)

Integrating strategic sensing with continuous customer discovery will improve organisational adaptability under conditions of technological and market uncertainty.

This proposition extends Dynamic Capabilities theory by proposing that sustained adaptation depends upon integrating external environmental sensing with ongoing empirical investigation of customer needs rather than treating these activities independently.

Proposition 6 (P6)

Product organisations evaluated primarily according to evidence quality and validated learning will demonstrate superior long-term innovation performance than organisations evaluated predominantly according to delivery-oriented performance metrics.

This final proposition reflects the broader shift proposed by HDNA from measuring innovation through implementation outputs towards evaluating innovation through organisational learning and improved strategic judgement.

Collectively, these propositions position HDNA as a theory-building framework rather than a prescriptive management methodology. They provide a structured basis for future empirical research examining how evidence-based governance influences innovation performance across organisations of different sizes, industries and levels of technological maturity.

3.9 Positioning the HDNA Framework Within Contemporary Innovation Theory

The HDNA Framework should not be interpreted as an alternative to existing innovation methodologies. Rather, its contribution lies in integrating complementary theoretical traditions into a coherent architecture for evidence-based organisational decision-making. Whereas most contemporary approaches optimise particular aspects of product discovery, HDNA explains how these diverse activities collectively contribute to reducing uncertainty throughout the innovation process.

Agile software development transformed the efficiency and responsiveness of software delivery but largely assumes that organisations have already identified the correct problems to solve. Lean Startup introduced validated learning through rapid experimentation but provides comparatively limited guidance regarding strategic capability or organisational governance. Customer Development strengthened customer problem discovery yet offers relatively little explanation of how customer insight should inform broader organisational decision-making. Design Thinking and Human-Centred Design significantly advanced empathy, creativity and user experience but remain primarily concerned with understanding customer needs rather than integrating that understanding with strategic resource allocation.

Similarly, Continuous Discovery institutionalises ongoing customer engagement throughout product development, yet focuses principally on customer learning rather than explaining how customer evidence should influence investment governance, organisational capability or long-term strategic adaptation. Stage-Gate methodologies provide valuable governance structures for managing innovation investment but have frequently been criticised for emphasising documentation, process compliance and administrative control rather than empirical learning.

HDNA seeks to synthesise these complementary perspectives rather than replace them. Each methodology contributes a distinctive capability that remains valuable within the broader discovery process. Agile contributes adaptive delivery; Lean Startup contributes experimentation; Customer Development strengthens opportunity validation; Design Thinking and Human-Centred Design enhance customer understanding; Continuous Discovery institutionalises learning; organisational learning theory explains knowledge creation; Dynamic Capabilities theory provides a strategic perspective on adaptation; and evidence-based management establishes principles for rational decision-making. HDNA integrates these contributions within a unified evidence-governed architecture that explicitly links discovery activities to organisational investment decisions.

This integrative positioning represents the principal theoretical contribution of the framework. Rather than introducing entirely new discovery techniques, HDNA provides an explanatory structure that demonstrates how established methods collectively function as components of a broader organisational learning system. Innovation therefore becomes less concerned with managing projects than with managing the progressive accumulation of reliable knowledge.

Importantly, the emergence of generative artificial intelligence further strengthens the relevance of this perspective. As argued throughout this dissertation, AI is rapidly reducing the cost, time and expertise required to develop software. Consequently, implementation is becoming increasingly commoditised, while customer understanding, strategic judgement and organisational learning become progressively more valuable sources of competitive differentiation. Existing innovation methodologies were largely developed under conditions in which software production represented the dominant organisational constraint. HDNA responds to this changing context by repositioning evidence generation as the primary objective of early-stage innovation.

Accordingly, the framework proposes a shift in the unit of organisational value creation. Rather than treating software as the principal output of product discovery, HDNA conceptualises validated knowledge as the critical organisational asset produced during early innovation. Software becomes the consequence of effective learning rather than the measure of innovation success itself.

This perspective concludes the theoretical development of the HDNA Framework and provides the conceptual foundation for the discussion that follows. The next chapter considers the broader implications of this evidence-based approach for product management, organisational learning and digital innovation in an era increasingly shaped by artificial intelligence.

4. Discussion: Reframing Product Discovery in the Age of Artificial Intelligence

4.1 From Building Products to Building Organisational Knowledge

The preceding chapters argued that the principal challenge confronting contemporary digital product development is no longer the efficient production of software but the effective reduction of uncertainty before substantial implementation resources are committed. Although advances in Agile development, DevOps and, more recently, generative artificial intelligence have transformed software engineering, these developments have not resolved the more fundamental question confronting organisations: how to determine which products are worth building in the first place. The HDNA Framework was developed in response to this challenge by reconceptualising product discovery as an evidence-governed process of organisational learning rather than a preliminary stage preceding software delivery.

This shift represents a significant departure from many traditional approaches to product management. Historically, product development has often been evaluated through implementation outcomes, including feature completion, delivery velocity, adherence to roadmaps and successful project execution. Such measures remain valuable indicators of operational performance; however, they provide limited insight into whether organisations are systematically improving their understanding of customers, markets and strategic opportunities. As argued throughout this dissertation, successful execution cannot compensate for pursuing opportunities that were poorly understood from the outset. Consequently, implementation efficiency and innovation effectiveness should be regarded as related but conceptually distinct organisational capabilities.

The HDNA Framework proposes that early-stage product development should therefore be understood primarily as a process of organisational knowledge creation. This perspective extends the work of Nonaka and Takeuchi (1995), who argued that sustained innovation depends upon organisations' ability to continuously create, integrate and apply new knowledge rather than merely accumulate information. Within the context of digital product discovery, customer interviews, market analysis, experimentation, usability testing and strategic reflection generate diverse forms of evidence that collectively improve organisational understanding. The strategic value of these activities lies not in the individual artefacts they produce, but in their cumulative contribution to reducing uncertainty and strengthening future decision-making.

Viewing discovery through the lens of organisational learning also provides a more robust explanation of why many innovation initiatives fail despite employing recognised product management methodologies. Existing approaches frequently improve discrete activities within the discovery process—such as customer research, experimentation or iterative delivery—yet often provide comparatively limited guidance regarding how evidence generated through these activities should influence broader organisational governance. As a result, organisations may gather substantial customer insight while continuing to prioritise investment decisions according to executive intuition, historical commitments or political considerations. Discovery therefore becomes operationally sophisticated without becoming strategically influential.

The HDNA Framework attempts to address this disconnect by positioning evidence as the principal currency of organisational decision-making. Rather than treating customer research as a validation exercise conducted after strategic direction has already been established, the framework embeds evidence generation throughout the entire discovery process. Every phase contributes to an evolving body of organisational knowledge that informs subsequent decisions and remains open to revision as new evidence emerges. In this respect, product discovery is no longer conceptualised as an activity that precedes development; instead, it becomes the organisational capability through which strategic judgement is continuously refined.

This perspective aligns closely with theories of double-loop learning (Argyris & Schön, 1978). Traditional project governance frequently encourages organisations to improve execution while leaving underlying strategic assumptions unchallenged. Teams become increasingly effective at delivering predetermined objectives, even when those objectives are themselves based on inaccurate assumptions regarding customer needs or market opportunities. Double-loop learning, by contrast, requires organisations to question the assumptions informing strategic decisions rather than merely optimising their implementation. The structured learning loops incorporated within HDNA operationalise this principle by explicitly requiring assumptions to be revisited whenever new evidence contradicts existing beliefs.

The implications extend beyond individual product initiatives to organisational capability more broadly. Knowledge generated through discovery should not remain confined to individual project teams but should accumulate as a strategic organisational resource capable of informing future innovation. Customer insights, validated assumptions, failed experiments and competitive observations collectively contribute to an organisational memory that improves subsequent opportunity identification and resource allocation. In this sense, the framework conceptualises discovery not simply as a project activity but as a cumulative capability through which organisations progressively strengthen their capacity to innovate under conditions of uncertainty.

Accordingly, the principal contribution of HDNA is not the introduction of new discovery techniques but the reframing of product discovery itself. The framework proposes that the fundamental output of early-stage innovation is neither prototypes nor product specifications but validated organisational knowledge. Software remains the eventual manifestation of successful innovation; however, knowledge constitutes the strategic asset through which organisations determine whether software should be developed at all. This conceptual shift establishes the foundation for reconsidering the changing economics of software development in an era increasingly shaped by artificial intelligence.

4.2 The Economics of Software Are Changing

One of the central propositions advanced throughout this dissertation is that the economics of digital product development are undergoing a fundamental structural transformation. Historically, software development was constrained by the scarcity of engineering expertise, lengthy development cycles and substantial implementation costs. Under these conditions, organisations derived competitive advantage largely through superior technical capability and the efficient execution of complex software projects. Consequently, much of the management literature focused on improving software delivery by reducing development risk, increasing productivity and enhancing implementation efficiency through methodologies such as Agile development, DevOps and continuous integration.

The emergence of generative artificial intelligence represents a significant departure from this historical context. AI-assisted programming, automated testing, interface generation and increasingly sophisticated software engineering agents substantially reduce the time, expertise and financial resources required to transform ideas into functioning software. While these technologies do not eliminate the need for skilled engineering, they reduce the marginal cost of implementation and expand the number of organisations capable of developing sophisticated digital products. Software production, once a major organisational constraint, is progressively becoming more accessible and increasingly commoditised.

This transformation has important strategic implications that extend beyond improvements in operational productivity. Much of the current literature evaluates AI primarily through its impact on software engineering performance, documenting gains in programming efficiency, documentation generation and knowledge-intensive work. Although these benefits are substantial, they represent only one dimension of AI's organisational significance. The more profound consequence is that AI alters the distribution of scarcity within digital innovation.

As implementation becomes progressively less constrained, competitive differentiation shifts towards capabilities that remain difficult to automate. The critical organisational challenge is no longer producing software efficiently but determining which opportunities justify investment. Customer understanding, strategic judgement, organisational learning and evidence-based decision-making therefore become increasingly valuable because they govern how organisations allocate resources within an environment where the capacity to build exceeds the capacity to choose wisely.

This argument extends Simon's (1971) observation that abundance creates new forms of scarcity. Just as information abundance transformed attention into a scarce organisational resource, AI-driven software abundance transforms strategic judgement into the primary constraint governing innovation. Organisations are now capable of generating substantially more ideas, prototypes and potential solutions than they can realistically evaluate or commercialise. Consequently, the limiting factor within innovation increasingly becomes the quality of organisational decision-making rather than the speed of implementation.

The implications for product management are considerable. Traditional product organisations have frequently been structured around the assumption that engineering capacity represents the principal bottleneck within innovation. Discovery activities therefore often functioned primarily as mechanisms for reducing development risk before expensive implementation commenced. As AI reduces implementation costs, however, the relative value of early-stage discovery increases. Investing additional effort in understanding customers, evaluating opportunities and testing assumptions becomes economically rational because the cost of selecting the wrong opportunity now exceeds the cost of producing software itself.

This shift also challenges conventional measures of innovation performance. Organisations that continue evaluating product teams primarily through delivery-oriented metrics—including feature throughput, release frequency or roadmap completion—may inadvertently reinforce behaviours that are increasingly misaligned with the emerging economics of software development. When implementation becomes comparatively inexpensive, producing more software does not necessarily create more value. Instead, value increasingly depends upon selecting opportunities capable of generating meaningful customer and organisational outcomes.

The HDNA Framework responds directly to this transformation by repositioning product discovery as the principal mechanism through which organisations allocate scarce strategic attention. Rather than treating discovery as an activity that supports delivery, the framework argues that delivery increasingly supports discovery. Software becomes a vehicle through which assumptions are investigated and customer value is realised rather than the primary objective of innovation itself. As AI continues to reduce implementation barriers, the strategic importance of evidence generation, organisational learning and disciplined opportunity selection is therefore likely to increase rather than diminish.

4.3 Artificial Intelligence Does Not Eliminate Uncertainty

Public discourse surrounding generative artificial intelligence frequently assumes that faster software development will naturally produce better innovation outcomes. Implicit within this assumption is the belief that reducing implementation effort also reduces innovation risk. The conceptual argument advanced throughout this dissertation challenges this interpretation. Although AI substantially accelerates software production, it does comparatively little to reduce the uncertainty that fundamentally characterises innovation.

Innovation uncertainty differs qualitatively from implementation uncertainty. Implementation concerns whether a proposed solution can be developed efficiently using available technologies and organisational resources. Innovation uncertainty, by contrast, concerns whether the underlying opportunity is worth pursuing at all. It encompasses questions relating to customer needs, market timing, competitive dynamics, strategic fit and organisational capability—questions that cannot be resolved solely through computational analysis or software generation.

Generative AI demonstrates considerable proficiency in synthesising information, generating software artefacts, producing design alternatives and assisting with analytical tasks. Nevertheless, these capabilities operate primarily upon existing data and probabilistic inference rather than direct engagement with evolving organisational and customer contexts. AI can recommend potential solutions, but it cannot independently determine whether a customer problem is sufficiently significant to justify investment, whether a proposed opportunity aligns with an organisation's long-term strategic objectives or whether changing market conditions fundamentally alter the attractiveness of a particular innovation. Such judgements remain dependent upon empirical observation, organisational interpretation and contextual reasoning.

Paradoxically, AI may increase rather than decrease the strategic importance of product discovery. Lower implementation costs reduce the financial penalty associated with experimentation while simultaneously increasing the number of feasible opportunities competing for organisational attention. As the number of possible products expands, selecting among them becomes progressively more complex. Organisations therefore require stronger mechanisms for evaluating evidence, prioritising opportunities and avoiding investment in superficially attractive but strategically weak initiatives.

This phenomenon may be conceptualised as the AI Acceleration Paradox: as the cost of building software decreases, the cost of deciding what to build increases. Under previous technological conditions, implementation constraints naturally limited the number of initiatives organisations could pursue. AI removes many of these constraints, creating an environment characterised not by scarcity of technical capability but by abundance of potential opportunities. The organisational challenge therefore shifts from production towards selection.

This shift has important implications for managerial decision-making. Faster implementation may create an illusion of increased certainty by enabling organisations to prototype and deploy products more rapidly. However, rapid execution does not substitute for systematic learning. Building a product quickly does not demonstrate that the product addresses an important customer need, creates sustainable competitive advantage or represents an appropriate allocation of organisational resources. Without structured evidence generation, accelerated implementation may simply enable organisations to pursue poor strategic decisions at greater speed.

Accordingly, the contribution of artificial intelligence should be understood primarily as augmenting organisational learning rather than replacing it. AI possesses considerable potential to strengthen discovery activities by synthesising qualitative data, identifying emerging behavioural patterns, supporting market intelligence and accelerating experimental design. Used in this way, AI expands organisational capacity to generate and interpret evidence. However, the interpretation of that evidence, the evaluation of competing strategic alternatives and the governance of investment decisions remain fundamentally human and organisational responsibilities.

The HDNA Framework reflects this distinction by positioning AI as a capability that enhances evidence generation while preserving the central role of empirical validation. Customer interviews, behavioural observation, experimentation and strategic reflection remain indispensable because they provide forms of contextual knowledge that cannot be inferred solely from computational models. AI therefore strengthens the discovery process not by replacing human judgement but by enabling organisations to learn more rapidly and more systematically.

Ultimately, the significance of AI lies less in automating innovation than in reshaping where innovation occurs. As implementation becomes increasingly automated, sustainable competitive advantage depends progressively upon those organisational capabilities that remain resistant to automation: curiosity, critical reasoning, contextual understanding and disciplined evidence-based decision-making.

4.4 From Product Discovery to Evidence-Governed Innovation

The arguments developed throughout this dissertation collectively suggest that contemporary product discovery requires a broader conceptual foundation than that provided by existing innovation methodologies. While approaches such as Lean Startup, Design Thinking, Customer Development and Continuous Discovery have substantially improved organisations' ability to investigate customer needs and evaluate emerging opportunities, they remain primarily concerned with improving discovery activities rather than explaining how discovery should be governed as an organisational capability.

The HDNA Framework proposes that this distinction is increasingly important. Discovery should not be understood as an isolated phase preceding software development, nor as a collection of customer research techniques performed independently of strategic decision-making. Rather, it should be conceptualised as an evidence-governed organisational process through which uncertainty is systematically reduced before substantial implementation commitments are made.

This perspective redefines the relationship between discovery and governance. Traditional governance frameworks frequently evaluate innovation through administrative artefacts such as business cases, stage-gate documentation, project milestones and delivery metrics. Although these mechanisms provide useful organisational control, they often measure compliance with planned activities rather than improvements in organisational understanding. As a consequence, projects may continue because documentation has been completed rather than because evidence justifies continued investment.

Evidence-governed innovation reverses this logic. Within the HDNA Framework, governance becomes an exercise in evaluating knowledge rather than monitoring execution. Investment decisions are informed by the credibility, consistency and sufficiency of the evidence generated throughout discovery rather than by the completion of predetermined deliverables. Progress therefore reflects increasing confidence in organisational judgement rather than increasing commitment to implementation.

This reconceptualisation also strengthens the relationship between product management and organisational learning theory. Discovery activities—including customer research, experimentation, usability evaluation, competitive analysis and strategic reflection—are no longer regarded as isolated techniques but as interconnected mechanisms through which organisational knowledge accumulates over time. Their collective value lies not merely in improving individual products but in enhancing the organisation's long-term capacity to recognise, evaluate and exploit emerging opportunities.

Importantly, evidence-governed innovation encourages a different organisational culture from many traditional product environments. Instead of rewarding certainty, it rewards disciplined inquiry. Rather than viewing invalidated assumptions as evidence of poor performance, organisations recognise them as valuable learning outcomes that prevent substantially greater losses during later stages of development. This cultural shift aligns closely with research on psychological safety and learning organisations, where constructive disagreement, reflective practice and openness to contradictory evidence are recognised as prerequisites for sustained adaptation.

The governance implications are particularly significant in the context of artificial intelligence. As implementation barriers continue to decline, organisations capable of systematically evaluating evidence are likely to outperform those that simply automate software production. AI may substantially improve execution, but it cannot determine which opportunities align with organisational purpose, customer value and long-term strategic capability. These decisions require governance systems capable of integrating diverse forms of evidence into coherent strategic judgement.

Accordingly, the HDNA Framework extends the concept of product discovery beyond its traditional operational boundaries. It positions discovery as the central organisational capability through which strategic decisions are informed, assumptions are challenged and innovation resources are allocated. In doing so, the framework reframes digital innovation as a continuous process of evidence accumulation, interpretation and organisational learning rather than a sequence of activities culminating in software delivery.

Viewed in this way, evidence itself becomes the principal governance mechanism. The purpose of discovery is not simply to understand customers or generate product ideas but to provide decision-makers with increasingly reliable knowledge upon which responsible investment decisions can be based. This represents the central conceptual contribution of the HDNA Framework and provides the foundation for considering its broader theoretical contributions, practical implications and future research opportunities in the remainder of this chapter.

4.5 Integrating Strategy and Discovery

One limitation identified across existing innovation methodologies is the separation between strategic planning and customer discovery.

Corporate strategy often remains an executive activity conducted periodically, while product discovery occurs operationally within individual product teams. This separation creates a risk that discovery activities optimise local customer problems without contributing to broader organisational objectives.

HDNA attempts to bridge this divide by beginning with strategic context rather than immediate solution ideation.

The Envisioning phase explicitly examines organisational capabilities, market dynamics, technological change and strategic intent before customer hypotheses are formulated. Consequently, discovery activities become aligned with broader organisational priorities while retaining flexibility to revise assumptions as evidence accumulates.

This integration reflects Teece's (2018) argument that sensing opportunities cannot be separated from organisational capabilities. Valuable opportunities emerge not only because customers possess unmet needs but because particular organisations possess distinctive capabilities enabling them to address those needs more effectively than competitors.

Thus, HDNA positions product discovery as the intersection of three forms of evidence:

• evidence regarding customers,

• evidence regarding markets,

• evidence regarding organisational capability.

Only where these domains overlap does sustained competitive advantage become plausible.

4.6 Limitations of the HDNA Framework

Although the HDNA framework integrates multiple theoretical perspectives, several limitations should be acknowledged.

First, the framework remains conceptual. While informed by established theories of innovation, organisational learning and product management, empirical evidence demonstrating its effectiveness across diverse organisational contexts is currently limited. Future research should therefore evaluate the framework through longitudinal case studies, controlled field experiments and comparative organisational analyses.

Second, HDNA assumes that organisations possess sufficient autonomy and resources to conduct iterative discovery activities. Highly regulated industries, safety-critical systems or public-sector organisations may face constraints limiting extensive experimentation. In such contexts, the framework may require adaptation to accommodate regulatory requirements and governance structures.

Third, successful implementation depends heavily upon organisational culture. Evidence-based decision-making requires leaders willing to revise assumptions, terminate weak initiatives and tolerate uncertainty. Organisations characterised by rigid hierarchies or strong political incentives may struggle to realise the full benefits of the framework despite adopting its formal processes.

Finally, the framework currently emphasises product discovery prior to implementation. Future extensions should examine how evidence generation continues throughout product scaling, platform evolution and ecosystem development.

4.7 Contributions to Theory

The HDNA framework contributes to the literature in three principal ways.

First, it synthesises previously fragmented approaches—including Lean Startup, Customer Development, Design Thinking, Continuous Discovery, Agile development and evidence-based management—into a unified conceptual architecture. Rather than introducing new methods, the framework demonstrates how existing approaches can be integrated around a common logic of uncertainty reduction.

Second, the framework explicitly incorporates generative AI into theories of product discovery. Existing innovation literature has largely examined AI as a productivity-enhancing technology. This paper argues instead that AI changes the strategic locus of competitive advantage by reducing implementation costs and increasing the relative importance of discovery.

Third, HDNA reframes product discovery as an organisational learning capability rather than a preliminary project phase. This perspective connects product management with broader theories of knowledge creation, dynamic capabilities and organisational adaptation, providing a richer theoretical foundation for future empirical investigation.

5. Managerial Implications

5.1 Rebalancing Discovery and Delivery

A central implication of this research is that organisations should reconsider the balance between product discovery and product delivery. Over the past two decades, significant investment has been directed towards improving software engineering capabilities through Agile methodologies, DevOps and continuous integration. These approaches have substantially increased the efficiency and reliability of software delivery. However, comparatively less attention has been devoted to institutionalising evidence-based discovery processes that ensure organisations are solving the right problems before investing in implementation.

This imbalance reflects an earlier technological environment in which software development represented the primary organisational constraint. Advances in cloud infrastructure, platform engineering, low-code technologies and generative artificial intelligence have significantly reduced the cost, time and complexity associated with software production. As implementation becomes increasingly commoditised, the strategic bottleneck shifts towards identifying valuable opportunities, validating customer needs and reducing uncertainty before committing resources.

Managers should therefore treat product discovery as a core organisational capability rather than an optional preliminary activity. Discovery deserves dedicated investment, governance and performance management because it directly influences the quality of strategic decision-making. Rather than viewing discovery as an activity that delays delivery, organisations should recognise that systematic uncertainty reduction early in the product lifecycle frequently prevents substantially greater costs associated with redesign, product repositioning and market failure later in development. Investment in learning ultimately improves both efficiency and long-term innovation performance.

5.2 Redefining Success Metrics

The findings also suggest that organisations should reconsider how they measure product team performance. Many organisations continue to evaluate success primarily through operational indicators that emphasise execution efficiency, including delivery speed, release frequency, roadmap adherence and development throughput. While these measures remain valuable for assessing delivery capability, they provide relatively little insight into whether teams are generating customer value or making better strategic decisions.

The HDNA framework proposes that traditional delivery metrics should be complemented by measures that assess organisational learning and evidence generation. Product teams should be recognised for reducing uncertainty, validating assumptions and improving the quality of decisions rather than simply producing more features. Metrics might therefore focus on the rate at which critical assumptions are validated, the effectiveness of experiments, the quality of customer insights generated and the speed with which teams incorporate new evidence into strategic decisions.

Adopting learning-oriented metrics encourages behaviours that prioritise adaptability, curiosity and evidence-based decision-making over output alone. This shift aligns organisational incentives with long-term innovation outcomes by recognising that successful products emerge from continuous learning rather than simply efficient execution.

5.3 The Evolving Role of Product Leadership

The findings presented in this paper also imply an important evolution in the role of product leadership. Historically, product managers frequently acted as coordinators responsible for gathering requirements, managing roadmaps and facilitating communication between business stakeholders and development teams. While these responsibilities remain important, contemporary product management increasingly requires leaders who can navigate uncertainty, interpret evidence and enable organisational learning.

Within the HDNA framework, effective product leadership is characterised less by possessing the correct answers than by asking increasingly insightful questions. Product leaders must continually examine the assumptions underpinning strategic decisions, identify areas of greatest uncertainty, determine what evidence would justify changing direction and assess whether new learning supports or contradicts existing beliefs. Decisions about continuing, adapting or abandoning product initiatives should be informed by evidence rather than commitment to predetermined plans.

This perspective reflects broader developments in adaptive leadership theory, which views leadership as creating the conditions for collective learning instead of directing fixed solutions. Consequently, successful product leaders require capabilities extending beyond technical expertise to include systems thinking, facilitation, behavioural decision-making, critical evaluation of evidence and the ability to foster productive collaboration across organisational boundaries.

5.4 Artificial Intelligence as a Discovery Partner

Much of the current discussion surrounding artificial intelligence focuses on its ability to automate software development activities such as programming, testing and documentation. The HDNA framework suggests a considerably broader role. Rather than simply accelerating delivery, AI has the potential to enhance product discovery by supporting the generation, synthesis and evaluation of evidence throughout the innovation process.

Generative AI can assist organisations by analysing large volumes of customer feedback, identifying emerging behavioural patterns, generating alternative product hypotheses, supporting competitive intelligence, summarising qualitative research, mapping opportunities, conducting scenario analysis and producing low-fidelity prototypes that can be rapidly tested with users. These capabilities enable teams to explore a wider range of possibilities while reducing the effort required to process complex information.

Nevertheless, artificial intelligence should be viewed as an augmentation technology rather than a replacement for customer engagement or strategic judgement. Large language models infer patterns from existing information but cannot independently determine whether a proposed opportunity represents a genuine unmet customer need or accurately interpret the contextual factors influencing customer behaviour. Human expertise therefore remains essential for evaluating evidence, understanding organisational context and making strategic trade-offs. The objective is not automated product strategy but enhanced organisational learning supported by intelligent analytical tools.

5.5 Building Organisational Learning Capabilities

Implementing evidence-based product discovery requires organisational change that extends well beyond individual product teams. Sustainable improvement depends upon creating an environment in which learning becomes an embedded organisational capability rather than an isolated project activity.

Learning-oriented organisations cultivate psychological safety so that employees feel confident challenging assumptions, presenting contradictory evidence and acknowledging uncertainty without fear of negative consequences. They make strategic assumptions explicit, ensuring that hypotheses are documented, tested and reviewed rather than remaining hidden within planning documents or informal discussions. Cross-functional collaboration is also essential because meaningful discovery depends upon integrating perspectives from product management, engineering, design, marketing, sales and executive leadership.

Governance processes should likewise become increasingly iterative, replacing infrequent stage-gate approvals with regular evidence reviews that enable decisions to evolve as new information emerges. Equally important is the development of reflective practice, whereby teams evaluate not only product outcomes but also the quality of the decision-making processes that produced those outcomes. By routinely examining how evidence was gathered, interpreted and applied, organisations strengthen their capacity for continuous improvement and organisational learning.

Collectively, these capabilities transform product discovery from a collection of isolated activities into an enduring organisational competence that supports long-term innovation and strategic adaptability.

5.6 Governance Through Evidence Rather Than Certainty

Traditional governance models often assume that increasing project maturity should correspond with increasing certainty. In practice, however, digital innovation rarely follows predictable trajectories. New technologies, evolving customer expectations and changing competitive environments mean that uncertainty frequently persists throughout the product lifecycle.

The HDNA framework proposes an alternative governance approach in which progression is determined by the quality of evidence rather than by project momentum or adherence to predetermined plans. At each stage of discovery, organisations should evaluate which assumptions have been validated, what uncertainties remain unresolved, what evidence supports the current strategic direction and whether continued investment is justified given the available information.

This evidence-based governance model encourages adaptive decision-making while reducing the risk of escalation of commitment, whereby organisations continue investing in initiatives despite mounting evidence that assumptions are incorrect. Governance therefore shifts from asking whether a project remains on schedule to determining whether sufficient evidence exists to justify further investment.

5.7 Implications for Digital Transformation

Many organisations continue to interpret digital transformation primarily as a technological initiative involving cloud migration, automation or artificial intelligence adoption. The findings presented in this research suggest a broader perspective. Successful digital transformation depends not only on technological capability but also on transforming how organisations make strategic decisions under conditions of uncertainty.

Technology undoubtedly accelerates execution and increases operational efficiency, yet it cannot compensate for poorly informed strategic choices. Organisations that invest heavily in advanced technologies while neglecting discovery capabilities risk accelerating the delivery of products that have not been adequately validated. As implementation becomes faster and less expensive, the quality of organisational learning becomes an increasingly important determinant of competitive advantage.

Digital maturity should therefore be evaluated not solely by technological sophistication but also by an organisation's ability to generate, interpret and apply customer evidence in support of strategic decision-making. Organisations capable of integrating technological capability with evidence-based discovery are likely to achieve more sustainable innovation outcomes than those focusing exclusively on delivery efficiency.

5.8 A Maturity Model for Evidence-Based Product Discovery

The HDNA framework also provides the foundation for a maturity model that enables organisations to assess the development of their discovery capabilities. Rather than representing a rigid sequence of stages, the model should be understood as a reflective framework that helps organisations evaluate how systematically they manage uncertainty and incorporate evidence into product decisions.

At lower levels of maturity, organisations tend to rely heavily on assumptions, intuition and delivery-focused planning, with limited validation of customer needs before implementation. As maturity increases, discovery activities become progressively more structured, assumptions are explicitly documented and experimentation becomes a routine component of product development. More advanced organisations establish governance processes that prioritise evidence over certainty, integrate learning across functional boundaries and continuously refine strategic decisions based on emerging insights.

The highest levels of maturity are characterised by discovery becoming an embedded organisational capability rather than an isolated product management practice. In these organisations, learning is treated as a strategic asset, evidence consistently guides investment decisions and leadership actively cultivates a culture of experimentation, reflection and continuous improvement. While organisations may progress through these capabilities at different rates depending on their context, advancing towards evidence-governed decision-making is likely to improve resilience and innovation performance in increasingly uncertain digital markets.

5.9 Summary

The managerial implications presented in this chapter reinforce the central argument of this research: competitive advantage in digital product development increasingly depends on an organisation's ability to reduce uncertainty through systematic learning rather than simply accelerating software delivery. As technological barriers continue to diminish, the capacity to generate, interpret and act upon evidence becomes an increasingly important source of strategic differentiation. Organisations that invest in discovery capabilities, evidence-based governance, adaptive leadership and organisational learning will be better positioned to identify valuable opportunities, respond to changing market conditions and deliver products that create sustained customer value.

6. Conclusion

6.1 Revisiting the Research Questions

This paper set out to investigate three interrelated research questions concerning contemporary digital product development.

The first research question asked:

Why do digital products continue to fail despite substantial improvements in software development capabilities?

The literature reviewed throughout this paper suggests that improvements in software engineering have not been matched by equivalent advances in product discovery. Agile methodologies, DevOps and AI-assisted development have dramatically increased organisations' ability to build software rapidly and reliably. However, these advances primarily optimise implementation rather than the identification and validation of valuable customer problems.

Consequently, many organisations continue to experience product failure not because software is poorly engineered, but because insufficient evidence exists that the proposed solution addresses a meaningful customer need. The central challenge therefore shifts from technical execution towards organisational learning and evidence-based decision-making.

The second research question considered:

How has generative artificial intelligence changed the relative importance of product discovery compared with software delivery?

The analysis presented in this paper argues that AI fundamentally alters the economics of digital innovation. By reducing the cost and time required to develop software, AI lowers barriers to implementation while simultaneously increasing the strategic importance of selecting appropriate opportunities.

As software generation becomes increasingly commoditised, sustainable competitive advantage is likely to depend less upon development capacity and more upon organisations' ability to understand customers, identify emerging opportunities and continuously generate evidence that informs strategic decisions.

Rather than diminishing the need for discovery, AI amplifies its importance.

The third research question examined:

How can organisations structure evidence generation throughout product discovery to improve strategic decision quality?

The HDNA framework provides one possible conceptual response.

By integrating strategic context, customer discovery, hypothesis validation and structured learning loops into a coherent decision architecture, the framework seeks to progressively reduce uncertainty before significant implementation commitments are made.

Importantly, HDNA does not propose replacing Agile, Lean Startup, Design Thinking or Continuous Discovery. Instead, it positions these approaches within a broader evidence-based model of organisational learning.

6.2 Theoretical Contributions

This paper makes three principal theoretical contributions.

First, it synthesises multiple streams of innovation research—including Agile software development, Lean Startup, Customer Development, Human-Centred Design, Continuous Discovery, evidence-based management and organisational learning—into a unified conceptual framework for digital product discovery.

While these traditions have developed largely independently, they share a common objective: reducing uncertainty through iterative learning. HDNA contributes by explicitly integrating these perspectives within a structured sequence of discovery activities connected through evidence review loops.

Second, the paper extends contemporary discussions regarding artificial intelligence in product development.

Much existing literature focuses on AI's capacity to improve software engineering productivity. This paper argues that AI should instead be viewed as changing the strategic distribution of organisational value creation.

As implementation becomes increasingly automated, customer understanding, strategic interpretation and evidence generation emerge as the primary sources of competitive differentiation.

Third, the framework contributes to product management scholarship by conceptualising discovery as an organisational capability rather than a project phase.

Rather than treating discovery as a collection of techniques, HDNA positions it as a continuous learning system embedded within organisational governance and strategic decision-making.

6.3 Practical Contributions

For practitioners, the framework suggests that successful digital innovation increasingly depends upon organisational capabilities that extend beyond software engineering excellence.

Managers should invest not only in improving delivery efficiency but also in strengthening evidence generation, customer discovery and strategic reflection.

The framework further recommends that organisations:

• document assumptions explicitly,

• validate hypotheses continuously,

• evaluate evidence before increasing investment,

• integrate discovery throughout product development,

• align governance with learning rather than output.

These recommendations become particularly relevant as AI accelerates development cycles and lowers implementation costs.

In rapidly evolving digital markets, organisations capable of learning faster than competitors may achieve greater long-term innovation performance than those capable merely of delivering software more efficiently.

6.4 Limitations

Several limitations should be recognised.

Most importantly, HDNA remains a conceptual framework.

Although grounded in established theories from innovation management, organisational learning and product development, the framework has not yet been subjected to large-scale empirical evaluation.

Consequently, the paper should be interpreted as theory development rather than theory confirmation.

Second, the framework has been developed primarily within the context of digital product innovation.

Its applicability to highly regulated sectors, public administration, healthcare, defence or safety-critical engineering requires further investigation.

Third, organisational culture represents an important moderating variable.

Evidence-based product discovery assumes leaders are willing to revise assumptions, terminate unsuccessful initiatives and embrace uncertainty.

Future research should therefore examine how leadership behaviour, organisational structure and governance influence successful implementation.

Finally, the paper has focused primarily on early-stage product discovery.

Additional research should investigate how evidence generation evolves during scaling, platform development and long-term product lifecycle management.

6.5 Future Research

The conceptual nature of HDNA creates several opportunities for future empirical research.

Longitudinal Case Studies

Future studies could examine organisations implementing HDNA over extended periods to understand how discovery practices influence innovation outcomes, decision quality and organisational learning.

Comparative Organisational Studies

Comparative research could evaluate organisations using HDNA against organisations employing Lean Startup, Design Thinking or conventional Agile product management.

Outcome variables might include:

• product-market fit,

• innovation performance,

• time-to-validation,

• customer adoption,

• development efficiency,

• investment efficiency.

Quantitative Validation

Future research could operationalise the propositions developed in this paper by constructing measurement instruments for:

• evidence quality,

• hypothesis maturity,

• uncertainty reduction,

• discovery capability,

• learning effectiveness.

Structural equation modelling could then examine relationships between discovery capability and innovation performance across multiple organisations.

AI-Augmented Product Discovery

An important emerging research area concerns AI-assisted discovery.

Future work should investigate:

• How effectively can AI synthesise qualitative customer evidence?

• Which discovery activities benefit most from AI augmentation?

• How should organisations combine AI-generated insights with human judgement?

• What governance mechanisms ensure responsible AI-supported product decisions?

These questions represent a promising intersection between product management, information systems and artificial intelligence research.

6.6 Final Remarks

Digital product development is entering a new era.

For much of the past three decades, organisations competed by improving the efficiency with which software could be designed, developed and delivered. Today, advances in artificial intelligence are rapidly reducing the cost of these activities, making technical implementation increasingly accessible and automated.

This paper argues that such developments fundamentally reshape where competitive advantage resides.

In an environment where building software becomes progressively easier, determining what should be built—and why—becomes the defining strategic challenge.

The HDNA framework proposes that this challenge is best addressed through evidence-based product discovery: a systematic process of transforming assumptions into validated knowledge before substantial implementation commitments are made.

Whether HDNA ultimately proves effective remains an empirical question.

However, the broader argument advanced in this paper is unlikely to diminish in relevance. As artificial intelligence continues to accelerate software production, organisations that cultivate superior learning capabilities, stronger customer understanding and more disciplined evidence generation are likely to outperform those relying primarily on faster implementation.

The future of digital innovation may therefore depend not on building more software, but on building better knowledge.

7. References

Argyris, C. and Schön, D.A. (1978) Organizational Learning: A Theory of Action Perspective. Reading, MA: Addison-Wesley.

Barney, J.B. (1991) ‘Firm resources and sustained competitive advantage’, Journal of Management, 17(1), pp. 99–120.

Beck, K., Beedle, M., van Bennekum, A., Cockburn, A., Cunningham, W., Fowler, M., Grenning, J., Highsmith, J., Hunt, A., Jeffries, R., Kern, J., Marick, B., Martin, R.C., Mellor, S., Schwaber, K., Sutherland, J. and Thomas, D. (2001) Manifesto for Agile Software Development.

Blank, S. (2013) The Four Steps to the Epiphany. 2nd edn. Pescadero, CA: K&S Ranch.

Blank, S. (2020) The Startup Owner's Manual. Pescadero, CA: K&S Ranch.

Brown, T. (2009) Change by Design. New York: Harper Business.

Brynjolfsson, E. and McAfee, A. (2017) Machine, Platform, Crowd. New York: W.W. Norton.

Brynjolfsson, E., Li, D. and Raymond, L.R. (2025) ‘Generative AI at work’, Quarterly Journal of Economics, 140(2), pp. 889–945.

Castellion, G. and Markham, S.K. (2013) ‘Perspective: New product failure rates: Influence of argumentum ad populum and self-interest’, Journal of Product Innovation Management, 30(5), pp. 976–979.

Cagan, M. (2020) Empowered. Hoboken: Wiley.

Cagan, M. (2020) Inspired: How to Create Tech Products Customers Love. 2nd edn. Hoboken: Wiley.

Christensen, C.M., Hall, T., Dillon, K. and Duncan, D.S. (2016) Competing Against Luck. New York: Harper Business.

Cooper, R.G. (2019) Winning at New Products. 5th edn. New York: Basic Books.

Cross, N. (2021) Design Thinking. London: Bloomsbury.

Denning, S. (2018) The Age of Agile. New York: AMACOM.

Dingsøyr, T., Nerur, S., Balijepally, V. and Moe, N.B. (2012) ‘A decade of agile methodologies: Towards explaining agile software development’, Journal of Systems and Software, 85(6), pp. 1213–1221.

Edmondson, A.C. (2019) The Fearless Organization. Hoboken: Wiley.

Forsgren, N., Humble, J. and Kim, G. (2018) Accelerate. Portland: IT Revolution.

Heifetz, R., Grashow, A. and Linsky, M. (2009) The Practice of Adaptive Leadership. Boston: Harvard Business Press.

ISO (2019) ISO 9241-210:2019 Ergonomics of Human-System Interaction – Part 210: Human-centred Design for Interactive Systems. Geneva: International Organization for Standardization.

Kahneman, D. (2011) Thinking, Fast and Slow. London: Penguin.

Liedtka, J. (2018) ‘Why design thinking works’, Harvard Business Review, 96(5), pp. 72–79.

March, J.G. (1991) ‘Exploration and exploitation in organizational learning’, Organization Science, 2(1), pp. 71–87.

McAfee, A. and Brynjolfsson, E. (2017) Machine, Platform, Crowd. New York: W.W. Norton.

Nonaka, I. and Takeuchi, H. (1995) The Knowledge-Creating Company. Oxford: Oxford University Press.

Norman, D.A. (2013) The Design of Everyday Things. Revised and Expanded edn. New York: Basic Books.

Noy, S. and Zhang, W. (2023) ‘Experimental evidence on the productivity effects of generative artificial intelligence’, Science, 381(6654), pp. 187–192.

OECD (2024) OECD Digital Economy Outlook 2024. Paris: OECD Publishing.

Perri, M. (2018) Escaping the Build Trap. Sebastopol, CA: O'Reilly Media.

Pfeffer, J. and Sutton, R.I. (2006) Hard Facts, Dangerous Half-Truths and Total Nonsense. Boston: Harvard Business School Press.

Porter, M.E. (1985) Competitive Advantage. New York: Free Press.

Ries, E. (2011) The Lean Startup. New York: Crown Business.

Royce, W.W. (1970) ‘Managing the development of large software systems’, Proceedings of IEEE WESCON, pp. 1–9.

Rousseau, D.M. (2006) ‘Is there such a thing as evidence-based management?’, Academy of Management Review, 31(2), pp. 256–269.

Schön, D.A. (1983) The Reflective Practitioner. New York: Basic Books.

Senge, P.M. (1990) The Fifth Discipline. New York: Doubleday.

Simon, H.A. (1971) ‘Designing organizations for an information-rich world’, in Greenberger, M. (ed.) Computers, Communications and the Public Interest. Baltimore: Johns Hopkins University Press, pp. 37–72.

Staw, B.M. (1976) ‘Knee-deep in the big muddy: A study of escalating commitment’, Organizational Behavior and Human Performance, 16(1), pp. 27–44.

Teece, D.J. (2018) ‘Business models and dynamic capabilities’, Long Range Planning, 51(1), pp. 40–49.

Torres, T. (2021) Continuous Discovery Habits. Sebastopol, CA: O'Reilly Media.

Ulwick, A.W. (2016) Jobs to Be Done: Theory to Practice. Denver: Strategyn Publishing.

Verganti, R. (2009) Design-Driven Innovation. Boston: Harvard Business Press.

Westerman, G., Bonnet, D. and McAfee, A. (2014) Leading Digital. Boston: Harvard Business Review Press.

Womack, J.P. and Jones, D.T. (2003) Lean Thinking. Revised edn. London: Simon & Schuster.

World Economic Forum (2025) The Future of Jobs Report 2025. Geneva: World Economic Forum.

Yin, R.K. (2018) Case Study Research and Applications. 6th edn. Thousand Oaks, CA: Sage.

Zott, C. and Amit, R. (2010) ‘Business model design: An activity system perspective’, Long Range Planning, 43(2–3), pp. 216–226.

Contact

Reach out via email for inquiries.

Email

Subscribe to newsletter

info@grcadvisory.ch

© 2025. All rights reserved.