Why Good Governance Kills Innovation — and What to Govern Instead
The initiative that clears every gate is not the most promising one; it is the one whose sponsor is most fluent in the language of false certainty.
The Question That Cannot Be Answered Honestly
Watch what happens when a genuinely new idea reaches the quarterly gate review. The board — the same board, sitting in the same room, that governs the delivery portfolio — works through its agenda, and the innovation initiative comes up between a systems upgrade that is running late and a compliance programme that cannot afford to. The questions put to it are the questions put to everything else: what is the three-year net present value, when does it reach break-even, and how confident are we in the plan. The people leading the idea have two options. They can tell the truth — we do not know yet; finding out is the entire point of the work — and watch it die for want of a number. Or they can manufacture a business case: a spreadsheet with a hockey-stick in the out-years, a discount rate applied to revenues nobody in the room can defend, a break-even date chosen not because anyone believes it but because it fits the funding cycle. Everyone present knows the second document is a work of fiction. And most of the time the fiction is what gets funded, because an honest answer is ungovernable and a dishonest one is not.
I have sat through enough of these reviews to be sure the fault is not with the people in the room. They are conscientious, and the discipline they are applying is real discipline. The trouble is that they are using an instrument built for one job to do a completely different one — and the instrument is working exactly as designed, which is what makes the damage so hard to see.
Two Different Problems, One Set of Instruments
Most of what an organisation does is delivery: building the thing it has already decided to build. Delivery is a problem of execution under known conditions. We know roughly what good looks like, we can estimate what it costs, and the honest measure of progress is variance against plan — are we where we said we would be, for the money we said it would take. Governance of delivery exists to answer one question well: are we executing the plan we approved. Stage-gates, business cases, milestone reporting, earned-value tracking — the whole apparatus is a certainty-management machine, and for delivery it is the right machine.
Innovation is not that problem. It is the work we take on precisely because we do not yet know what good looks like, cannot honestly estimate the cost, and have no plan worth measuring variance against. Its unit of progress is not “are we on plan” but “do we know more than we did.” Yet when the new idea reaches the gate we reach for the delivery instruments, because they are the only ones to hand and because they feel like rigour. We ask the discovery problem to answer the delivery question. It cannot — so it lies, or it dies.
Governance built to manage the absence of surprise will, turned on innovation, systematically fund the initiatives best at pretending the surprise is absent — and starve the ones honest enough to admit it is not.
This is the mechanism, and it is worth stating plainly because it inverts the usual complaint. The problem with governing innovation by business case is not that it is too strict. It is that it selects for the wrong thing. The initiative that clears every gate is not the most promising one; it is the one whose sponsor is most fluent in the language of false certainty. The one killed at gate two is not the weakest; more often it is the most honest about how much remains unknown. I have watched an idea turned down at a funding review for refusing to commit to a launch date, and watched a competitor carry the same idea into the market two years later. The gate had done its job perfectly. That was precisely the problem.
The Objection Worth Taking Seriously
There is a serious case on the other side, and it deserves better than the caricature reformers usually hand it. Financial discipline on innovation is not bureaucratic pedantry; it is the immune system that stops an organisation bleeding money into a hundred pet projects that never quite die. Loosen the demand for a business case and you do not get a garden of a thousand flowers — you get empire-building, sunk-cost escalation, and the well-connected executive whose favourite venture consumes budget for years on the strength of conviction alone. The discipline of “show me the numbers” exists because someone, somewhere, learned the hard way what happens without it. Anyone who has watched an unaccountable innovation unit burn through capital for a decade knows the objection is not theoretical.
The objection is right about the disease and wrong about the cure. The answer to undisciplined innovation spending is not the delivery business case, because the delivery business case does not in fact impose discipline on discovery — it imposes theatre. A three-year net present value on an idea whose core assumptions are untested is not a control; it is a number wearing the costume of a control, and everyone soon learns to produce it on demand. Real discipline for innovation exists. It is simply a different discipline, and a harder one — because it asks not “can you write me a convincing plan” but “what do you now know that you did not know last quarter, and what did it cost you to find out.”
Governing Learning Instead of Certainty
If the unit of progress in innovation is reduced uncertainty, then governance should fund and measure exactly that. This is not a relaxation of control. In several respects it is more exacting than the gate it replaces, because it will not accept a plan in place of evidence.
Three shifts do most of the work.
- Fund against evidence, not against plans. Release money in deliberately small tranches, each one buying the answer to a specific question rather than a slice of a predetermined build. The first tranche does not fund “phase one of the product”; it funds the cheapest test that could kill the idea outright. What earns the next tranche is not adherence to schedule but a genuine reduction in what is unknown — an assumption validated, one falsified, a risk retired. This is the logic of discovery-driven planning — managing the ratio of assumption to knowledge as the thing that actually matters — and of treating each increment of funding as the purchase of an option rather than the down-payment on a commitment.
- Measure the reduction of uncertainty, not variance against a baseline. The right question at the review is not “are you on plan” — there was never an honest plan — but “which of your critical assumptions have you tested, and what did you learn.” An initiative that has spent its money proving its central premise wrong has succeeded at the only thing that stage could succeed at: it has bought a cheap answer to an expensive question. Governance that cannot tell that apart from failure will punish the very behaviour it most needs to reward.
- Make ending an initiative a feature, not a verdict of blame. Under the business-case regime a cancelled initiative is a failure, so sponsors fight to keep the walking dead upright and the portfolio silts up with projects nobody believes in but nobody can be seen to stop. Governance of learning inverts this: a well-run initiative that cheaply establishes that an idea will not work is a good outcome, and the people who ran it should be redeployed with their standing enhanced, not dented. Cheap, early, blameless termination is the most valuable thing this kind of governance can produce, because it is what releases capital for the ideas that deserve it.
Set the two regimes side by side and the difference is not one of rigour but of what each is rigorous about.
| Dimension | Governance of delivery | Governance of discovery |
|---|---|---|
| Core question | Are we executing the plan? | Have we reduced what we don’t know? |
| Unit of funding | The approved project | The next decisive experiment |
| Measure of progress | Variance against baseline | Assumptions tested and retired |
| Role of failure | A deviation to be corrected | A cheap answer to be banked |
| Cadence | Periodic gates on the calendar | Frequent reviews on the evidence |
None of this argues for less oversight. A board governing discovery this way will meet more often, ask harder questions, and stop more initiatives than one running quarterly gates — it will simply stop them earlier, cheaper, and for the right reasons. The discipline has not been removed. It has been pointed at the thing that actually varies.
Why We Reach for the Wrong Instrument Anyway
If the alternative is this clear, why is the business case still the reflex? Partly because the delivery instruments are genuinely good at their own job, and success breeds over-reach: the machine that governs the ninety per cent of the portfolio that is delivery gets turned on the ten per cent that is not, simply because it is there and it is trusted. Partly because false certainty is more comfortable than honest uncertainty — a fictional net present value lets everyone in the room believe a decision has been taken on evidence, and that comfort is worth a great deal to people who are personally accountable for the money.
But the deepest reason is temperamental, not technical. Governing discovery well means a senior group has to sit with not-knowing: to approve spending against a question rather than a plan, to call an idea’s early death a good day’s work, to resist the almost physical pull toward a number that would make the discomfort go away. That is a matter of institutional temperament, and it is far scarcer than the technique. We are fluent in method; we are much less fluent in the composure that lets us govern honestly in the absence of answers.
“You cannot govern what you do not yet understand with instruments built to control what you already do. The first step is admitting they are two different jobs.”
The organisations that steward innovation well are not the ones with the most elaborate innovation frameworks. They are the ones that have learned to tell their two problems apart, and to keep the certainty-management machine away from the work whose entire value lies in the fact that the certainty is not there yet. It is a modest-sounding discipline. It is also, from what I have seen, the rarest one in the room.