Every technology project your company greenlights carries a number: an expected return on investment. The business case gets built, the spreadsheet gets blessed, and if the ratio looks right, the project moves forward. If it doesn’t, it dies in committee.
This is rational. It is also dangerously incomplete.
The standard ROI calculation treats every project as an isolated event; a self-contained transaction between capital spent and value received. But the most strategically valuable technology investments do something the spreadsheet never captures: they change the math on every project that comes after them. They don’t just generate a return. They generate an innovation dividend, a permanent reduction in the cost and risk of your next move, and the one after that, and the one after that.
Understanding why requires borrowing a concept from an unlikely source: theoretical biology.
The Adjacent Possible
In the 1990s, complexity theorist Stuart Kauffman introduced a concept he called the “Adjacent Possible” to describe how biological systems evolve. Steven Johnson later adopted the framework as the central organizing principle of his book Where Good Ideas Come From: The Natural History of Innovation, applying it to the entire arc of human invention. The idea is deceptively simple. At any given moment, a system (a cell, an organism, a company, an economy) can only reach configurations that are one step away from its current state. It cannot leap to some distant, optimal design; it can only explore the doors that are immediately available to it. But here is the critical insight: every time it steps through one of those doors, new doors appear that did not exist before.
The adjacent possible is not a fixed space. It expands with every step taken. In Kauffman’s original biological context, a molecule that couldn’t have formed in yesterday’s chemical environment becomes trivially accessible today, because yesterday’s reactions produced the precursors it needed. The entire history of biological innovation, from single-celled organisms to the Cambrian Explosion, follows this pattern: each small, viable step unlocking a cascade of previously impossible next steps.
Johnson showed that the same dynamic governs human innovation. Breakthrough technologies almost never appear out of thin air. They appear when prior innovations have assembled the right preconditions. The printing press required metallurgy, the screw press, oil-based inks, and cheap paper to already exist. The internet required packet switching, TCP/IP, and affordable personal computing to already be in place. Each of those precursor technologies was useful in its own right; none of them were built “for” the thing they eventually enabled. But without them, the adjacent possible never expanded far enough to include what came next.
We’ve incorporated this mindset into our Innovate Beyond Efficiency® framework, because this is exactly how the best technology strategies work inside companies. And it is exactly what standard project-by-project ROI calculations fails to see.
Two Kinds of Value · One Project
Consider two mid-market companies, both investing in a modern data platform. Company A treats it as a standalone initiative. The business case is built around operational reporting: faster dashboards, fewer manual Excel processes, maybe some improved inventory visibility. The ROI is defensible. The project gets funded, delivered, and the expected value materializes. Case closed.
Company B builds the same data platform, with the same operational reporting use case and the same defensible ROI. But the leadership team chose this project with a second lens in mind. They selected a platform architecture and data model that would also support machine learning workloads. They insisted on clean, governed data pipelines, not because the reporting project required that level of rigor, but because they knew what they wanted to do next. The project is delivered. The same operational value materializes. Same ROI on paper.
But Company B has done something Company A has not. It has expanded its adjacent possible. Six months later, when an AI-driven demand forecasting opportunity emerges, Company B’s “I” (the investment required) is a fraction of what Company A would face. Company A would need to go back and rebuild its data infrastructure, clean its data, establish governance, and retrain its teams before it could even begin. Company B has already done all of that; it was embedded in the project it already justified on other grounds.
The demand forecasting project that would have cost Company A $800,000 and taken eighteen months might cost Company B $200,000 and take four months. The return is comparable. The ROI is four times higher, not because the return is bigger, but because the investment is smaller.
That delta is the innovation dividend. It doesn’t show up in the original project’s business case. No spreadsheet captures it. But it is real, it is substantial, and over time it is the single biggest differentiator between companies that compound their technology advantages and companies that fight the same infrastructure battles over and over again.
Reducing the “I” in ROI
Most conversations about ROI focus on the numerator: the return. How much revenue will this generate? How much cost will it eliminate? How many hours will it save? These are important questions. But for companies thinking strategically about technology, the denominator is where the leverage lives.
Every well-chosen technology project should reduce the “I” in future projects’ ROI calculations. This happens through several mechanisms:
Capabilities that persist. A customer data platform built for marketing segmentation also serves as the foundation for personalized pricing, churn prediction, and customer lifetime value modeling. The investment in clean customer data, identity resolution, and integration architecture doesn’t need to be repeated; it carries forward.
Skills that compound. A team that delivers a successful cloud migration doesn’t just produce a migrated workload. It produces a team that now knows how to migrate workloads. The second migration is faster and cheaper. The tenth is almost routine. The organizational learning curve itself is a durable asset that reduces the “I” on every subsequent project.
Risk that diminishes. Uncertainty is expensive. It inflates timelines, demands larger contingency budgets, and slows decision-making. A company that has successfully delivered three iterative technology projects has something invaluable: institutional evidence that it can execute. The political cost of greenlighting project four drops considerably. The risk premium baked into the business case shrinks.
Architecture that opens doors. An API-first integration layer built for one system connection makes the next connection cheaper and the tenth connection almost free. A well-designed microservices architecture turns what would have been a six-month custom development effort into a configuration exercise. The right architectural choices don’t just solve today’s problem; they pre-solve tomorrow’s.
None of these benefits are speculative. They are observable, repeatedly, in companies that approach technology investment with strategic intent rather than project-by-project justification.
The Trap of the Isolated Business Case
If the innovation dividend is so powerful, why do most companies fail to capture it? Because their decision-making frameworks are not designed to see it.
The typical mid-market technology governance process evaluates each project on its individual merits. This is not unreasonable; you should not fund projects that cannot justify their own existence. But when individual justification is the only criterion, the result is a portfolio of disconnected investments that never compound.
Think of it like buying tools for a workshop, one at a time, with each purchase justified solely by the immediate job it addresses. You buy a drill because you need to hang shelves. You buy a saw because you need to cut boards. You buy a sander because you need to finish a table. Each purchase is rational in isolation. But if you never step back and ask “what kind of workshop am I building?” you end up with a cluttered garage full of single-purpose tools instead of a workshop that enables you to build anything.
The worst version of this trap is what many companies experience as “transformation fatigue”. Every few years, a major initiative is launched: ERP replacement, digital transformation, cloud migration, now AI. Each one is treated as a discrete event with its own timeline, budget, and success criteria. Each one disrupts the organization, consumes enormous resources, and delivers some portion of its promised value. And then, a few years later, the next wave arrives, and the company discovers that the last initiative left it no better prepared for this one than it was before. The “I” never goes down. The adjacent possible never expands. Each project starts from scratch.
Each Step Must Stand on Its Own
There is an important nuance here, and it is one that separates the adjacent possible framework from the kind of starry-eyed “visionary” thinking that has produced some of technology’s most expensive failures.
Every step must make sense on its own. Every project must have a defensible, standalone ROI. The innovation dividend is a bonus; it is never the justification.
This is the discipline that Kauffman’s framework demands. Biological evolution does not take steps that are harmful to the organism today in hopes that they will be useful in some future generation. Every mutation that persists must confer an advantage now, or at minimum not impose a fatal cost. The fact that it also opens new evolutionary possibilities is what makes the process so powerful, but the immediate viability is non-negotiable.
The same principle applies to technology strategy. A company should not build a data lake “because we might need it for AI someday” if the data lake itself delivers no current value. That is not strategic thinking; that is speculative spending dressed up in strategic language. But a company should absolutely choose the data platform architecture that supports both today’s reporting needs and tomorrow’s AI ambitions, when both options would satisfy the immediate requirement at comparable cost.
The art is in the selection, not the justification. You do not change the business case. You change which projects you choose to pursue, and how you architect the ones you approve, so that each investment leaves you better positioned for the next one.
The Compounding Effect
When this discipline is applied consistently, something remarkable happens. The cost of innovation itself begins to decline. Each project expands the adjacent possible, which surfaces new opportunities that would not have been visible before, and those opportunities arrive pre-discounted because the prior investments have already absorbed much of the cost.
This is the compounding effect that separates companies with genuine technology advantages from companies that simply spend a lot on technology. The former get more value per dollar with each successive investment. The latter face roughly constant (or even increasing) costs because they never build on what came before.
The parallel to compound interest is not accidental. In financial terms, the innovation dividend is the “interest on interest” that Albert Einstein (probably apocryphally) called the most powerful force in the universe. Each project’s dividend is modest on its own. But reinvested consistently across a portfolio of well-chosen investments over five or ten years, the cumulative effect is transformative.
A company that has spent $5 million on technology over five years through a sequence of strategically connected, individually justified projects will almost always be in a stronger competitive position than a company that spent $10 million on disconnected, large-scale initiatives over the same period. The first company has been compounding. The second has been spending.
Practical Application
How does a leadership team actually implement this? Not with a more complex spreadsheet. The innovation dividend is real, but attempting to quantify it precisely for each project is a fool’s errand; the interdependencies are too numerous and the timelines too uncertain.
Instead, the shift is qualitative and architectural. It requires three things:
A technology thesis. Leadership needs a clear, revisited-annually articulation of where the company’s technology capabilities are heading over the next three to five years. Not a detailed roadmap; a directional thesis. “We will become a data-driven organization that can personalize the customer experience at scale” is a thesis. It does not prescribe specific projects; it provides a filter for evaluating them.
Architectural intentionality. Every project should be evaluated not just on what it delivers, but on what it enables. When two approaches to solving the same problem cost roughly the same, choose the one that expands the adjacent possible. When an architectural choice would cost modestly more but would meaningfully reduce the “I” on two or three future initiatives, that premium is almost always worth paying.
Portfolio awareness. Stop evaluating projects in isolation. Evaluate them as a portfolio, the way a venture capitalist evaluates investments. Some projects are primarily about immediate return. Some are primarily about expanding optionality. The best ones do both. A governance process that can only see the first category will systematically underinvest in the second.
The Strategic Difference
The companies that consistently outperform their peers in technology-driven value creation are rarely the ones that spend the most. They are the ones that invest with the adjacent possible in mind; choosing projects that stand on their own merits today while quietly reducing the cost and risk of tomorrow’s opportunities.
This is what separates technology strategy from technology budgeting. Budgeting asks: “Does this project justify its cost?” Strategy asks that question too, and then asks a second one: “Does this project make the next one cheaper, faster, and more likely to succeed?”
When your answer to both questions is yes, you are not just generating a return. You are generating an innovation dividend. And unlike most returns, this one compounds.
