NOTE: This is Part 1 of a two-part op-ed series by Dr. Andrew Maxwell, the Bergeron Chair in Technology Entrepreneurship in the Lassonde School of Engineering at York University, looking at why innovation funding alone rarely produces economic impact. Part 2, which examines why diffusion, decision quality and governance determine the economic returns from innovation, will be published on February 25.

Innovation is widely understood as a central driver of long-run economic growth. Governments invest heavily in research, universities expand their innovation portfolios, and organizations track rising levels of inventive activity.

Yet across many economies, productivity growth has slowed and the economic returns to innovation investment appear increasingly uneven.

This tension is often framed as a puzzle: if innovation is essential to prosperity, why does more innovation not reliably translate into stronger economic outcomes?

The answer is not that innovation theory is wrong, nor that research investment lacks value. Rather, the answer lies in a gap between how innovation is conceptualized and how it is operationalized in policy and practice.

In particular, adoption is rarely treated as a central variable in how research is evaluated or how innovation decisions are made.

A substantial body of economic theory demonstrates how innovation expands productive possibilities by increasing the stock of knowledge, improving technologies, and enabling new forms of economic activity. These insights explain potential growth – the gains that would be achievable if economically relevant innovations were fully translated into productive use.

Observed economic outcomes, however, reflect realized growth. The difference between the two is not a rounding error. It is shaped by a sequence of transitions that determine whether innovative potential survives the journey from research to economic use.

When innovation policy focuses primarily on inputs – research funding, patent counts or startup formation – it implicitly assumes that these transitions will occur naturally. In practice, they often do not.

What is largely missing from this logic is any systematic consideration of whether, how, and by whom new knowledge will be taken up and used.

Where the innovation-to-impact chain breaks

Between research and economic impact lies a complex process involving invention, early use, broader adoption, diffusion and productive integration. At each stage, uncertainty increases, coordination becomes more difficult, and the costs of failure shift across actors.

Many promising ideas fail not because they lack technical merit, but because they encounter barriers to use: misalignment with organizational routines, lack of complementary capabilities, or insufficient incentives for early adopters. Others succeed locally but fail to diffuse in ways that reshape economic activity at scale.

At this stage, the challenge is not yet diffusion or scale, but the more basic question of whether innovations are taken up and used at all – a problem that reveals how little adoption is considered in research evaluation and early decision processes.

When these losses are not made explicit, innovation systems can appear highly active while producing limited transformation.

In the absence of a clear process model linking innovation to impact, policy responses often default to expanding research investment. This is understandable: research inputs are measurable, politically defensible and institutionally familiar.

But when downstream constraints dominate, additional research funding yields diminishing returns. The system generates more ideas without improving its ability to convert those ideas into productive use. Expectations rise, outcomes disappoint, and pressure builds for further intervention – often in the same direction.

The result is not a lack of innovation, but a misalignment between where effort is concentrated and where value is lost.

This raises a critical question: where does the absence of adoption thinking first become visible in the innovation process?

The first-user problem: Where the missing adoption logic first appears

If innovation systems fail to convert research into economic impact, the breakdown often begins earlier than expected. Long before diffusion stalls or productivity gaps widen, many innovations struggle to secure their first meaningful users.

This challenge – the first-user problem – is not the whole of the adoption issue. Rather, it is the earliest manifestation of a deeper omission: the fact that adoption is largely absent from how research is framed and how early innovation decisions are made.

Invention reduces technical uncertainty, but it does not eliminate uncertainty about use. Early adopters face risks that later users do not: unproven performance, unclear integration costs, uncertain returns and limited reference points.

These risks are asymmetric. Innovators benefit from experimentation and learning, but early users bear the costs. As a result, even innovations with strong long-term potential may struggle to find organizations willing to adopt them first.

This is not a failure of rationality. It reflects a coordination problem embedded in the innovation process – one that becomes acute precisely because adoption has not been treated as a core design and evaluation criterion upstream.

Innovation policy often treats pilots, demonstrations and proofs of concept as sufficient bridges between invention and use. These activities reduce some forms of uncertainty, but they rarely resolve the deeper challenges of organizational integration.

Meaningful first use requires innovations to be embedded in real workflows, decision structures and incentive systems. It exposes gaps in skills, processes and complementary assets that pilots are designed to avoid. As a result, many innovations appear successful in controlled settings but falter when confronted with the realities of practice.

When first use is superficial, the learning generated is limited – and the foundation for later adoption remains weak.

From demonstration to discovery

One way to address the first-user bottleneck is to shift attention from demonstrating technical feasibility to discovering user needs.

Processes commonly described as customer discovery or problem-solution exploration require researchers and innovators to identify and engage directly with potential first users early in the innovation journey. The purpose is not to “sell” a technology, but to understand real operational constraints, decision criteria, and sources of value from the user’s perspective.

This step is often absent from formal research evaluation and funding processes. Yet it is precisely through early engagement with prospective first users that innovators can learn whether a solution addresses a meaningful need, how it must be adapted for practical use, and what complementary changes are required for adoption.

In this sense, customer discovery is not a commercialization add-on; it is a mechanism for aligning research with use before the first adoption decision is even possible.

First users play a disproportionate role in shaping innovation trajectories. Their experiences generate the evidence that later adopters rely on. Their adaptations reveal how technologies interact with existing systems. Their successes and failures influence expectations about feasibility and value.

When first use is absent, delayed or poorly supported, innovations remain fragile. Diffusion becomes speculative rather than evidence-based, and subsequent adoption decisions are made with incomplete information.

In this sense, first use is not simply a milestone – it is a learning mechanism. It is also the point at which the absence of explicit adoption logic in research and early decision processes becomes most visible.

Which innovations reach first users is not random. Selection criteria, procurement rules and evaluation frameworks all shape which projects are advanced and which stall.

When these tools prioritize technical novelty or near-term performance, they may unintentionally favor innovations that are easy to demonstrate but difficult to integrate meaningfully. Innovations that require deeper organizational change – or that generate value through systemic effects – are often filtered out early, despite their long-term potential.

This bias has lasting consequences. Innovations that struggle at the first-use stage rarely recover later.

Why first use shapes everything that follows

Diffusion, scaling and economic impact depend critically on what happens at first use. When early adoption generates deep learning and credible evidence of value – often through structured engagement with first users during discovery and early use – later adoption becomes easier. When it does not, diffusion becomes slow, uneven or symbolic.

Addressing the first-user problem therefore requires more than encouraging experimentation. It requires aligning incentives, evaluation criteria and support mechanisms with the realities of early use – and, more fundamentally, recognizing adoption as a central dimension of innovation rather than a downstream afterthought.

This article has argued that innovation underperforms not because research is unproductive, but because adoption is largely absent from how research is evaluated and how early innovation decisions are made. The first-user problem reveals where this omission first constrains impact.

The second article in this two-part series turns to what happens when adoption does occur – why diffusion is uneven, why some forms of use generate far greater economic impact than others, and why how innovations are adopted matters as much as whether they spread.

Innovation creates possibilities. Impact depends on how those possibilities are taken up, used and learned from.

R$