Most organisations approach innovation investment the same way they approach operational investment. Justify every line item, minimise risk, maximise certainty before committing capital. It is a disciplined approach that works well for running a business, and it is almost perfectly designed to prevent innovation from happening.
The problem is not discipline or intent. It is the mental model. Operational investment logic assumes the outcome is reasonably predictable if the inputs are right. Innovation investment assumes the opposite: you do not yet know which of your assumptions are wrong, and the early capital exists to find out. Applying the first model to the second produces predictable results, and none of them are good.
Three ways it goes wrong
The failure modes are more varied than most post-mortems acknowledge. Overspending on the wrong things gets the most attention, while the other two are equally common and equally damaging.
The organisation that cannot spend. The budget exists and so does the mandate, but the decision-making machinery was built for a different purpose. Every deployment decision goes through the same approval process as a capital expenditure: business case, sign-off chain, risk committee, procurement. By the time approval comes through, the moment has passed or the team has lost momentum. The year ends with a fraction of the budget deployed and nothing meaningful built, an outcome identical to having no budget at all.
The organisation that drip-funds. Capital is released in amounts too small to build or learn anything meaningful. The team runs a workshop, produces a report, launches a small pilot, and nothing reaches the scale at which you could honestly assess whether it works. The organisation tells itself it is being prudent, when what it is actually doing is spending enough to feel like it is innovating without taking enough risk to find out whether it is.
The organisation that spends too fast on the wrong things. A large commitment is made before the problem is properly understood. An agency is hired, a team assembled, a launch event held, and eighteen months later there is an expensive product nobody wants, built on assumptions nobody validated. The problem was not the spending. It was the sequence.
What you invest in matters as much as how you invest
There is a fourth problem that sits underneath all three, and it is less about the amount of capital than its direction.
In practice, a significant proportion of innovation investment ends up flowing horizontally before the use cases that would justify it have been validated. A data platform, an API layer, an AI capability, a reusable component library. Infrastructure that theoretically enables future value but delivers nothing on its own and is almost impossible to evaluate because there is no direct business outcome to measure against. The logic is appealing: build the foundations and the products will follow. In practice, the foundations get built and the products never arrive, because the organisation has spent its energy and appetite on the infrastructure rather than the thing the infrastructure was supposed to enable.
The most common justification is shared capability: build it once and every team benefits. The problem is that “everyone will use this” is a projection made before anyone has validated what they actually need the capability for. By the time the platform is ready, the intended use cases have shifted, teams have found workarounds, and the specification turns out to have been written for a future state that never arrived. Shared infrastructure also creates shared dependencies, and the team that needs to move fast finds itself waiting on a roadmap governed by committee.
Investing in shared capability is simultaneously a strategic, financial, and technical decision. When it gets made as a purely technical call, which it often does, the business case goes unexamined until the money is spent.
Vertical investment starts from the opposite end. Pick a specific business problem and solve it end to end, even if the solution is narrow. A customer does something, the business benefits measurably, and you can evaluate whether it worked. Use that validated demand as the foundation for shared infrastructure, designed to serve proven use cases rather than a projected future state. Built that way, on real evidence and incremental need, shared capability is a sound strategic investment. Built the other way, it is horizontal investment with a more convincing rationale.
The practical consequence of defaulting to horizontal is that venture portfolio thinking becomes almost impossible to apply. Without a specific outcome to measure, there is no milestone to evaluate and no principled basis for deciding whether to continue or stop. You just keep building.
The venture portfolio model
The frame that works is closer to how venture investors think about early-stage capital: small initial bets, deployed deliberately, against a specific question you are trying to answer. Is this problem real, and does our approach to solving it have merit?
Capital comes in tranches, each conditional on what the previous tranche taught you. The decision at each stage is whether what you have learned justifies continued investment, and if it does not, you stop. This is not failure. It is the system working. Money spent learning that an idea does not work is not wasted, provided the learning was genuine and the decision to stop was made cleanly. The waste is in continuing past the point where the evidence stopped supporting the investment, which is what happens when governance is built around avoiding the embarrassment of stopping rather than the discipline of learning.
The budget that never moves
The organisation that cannot deploy its innovation budget deserves particular attention because it is underreported. It is easy to identify the programme that overspent. It is considerably harder to see the cost of capital that sat undeployed because the machinery could not move fast enough to use it, and the result is the same: nothing built, nothing learned.
The cultural signal, though, is different and in some ways more damaging. An organisation that consistently fails to deploy its innovation budget has built a governance process designed to prevent loss rather than enable learning, and that is a leadership problem before it is a process problem. The process reflects a belief, usually unspoken, that the safest outcome is no decision.
The leadership shift
The executives who handle innovation capital well share a quality that is harder to develop than it sounds: they have made peace with productive loss. They understand that not every bet will pay off, that some of the most valuable outcomes are the ones that tell you clearly what not to pursue, and that the discipline of stopping is as important as the discipline of starting.
That is a different relationship with uncertainty than most senior careers prepare you for. Building a track record in a large organisation typically rewards being right, managing risk downward, and delivering against plan, whereas innovation investment rewards learning fast, changing your mind when the evidence changes, and making the stop decision before it becomes expensive.
The organisations that build that capacity in their leadership, not just their innovation teams, are the ones that compound over time. The ones that do not keep rebuilding the same programme under a different name.