Your Innovation Programme Didn’t Fail. Your Organisation Ate It.
The people were good. The idea held together. The budget was signed off in the room, with witnesses. None of it mattered, because your organisation runs on the same machinery that keeps the core business safe and boring, and that machinery cannot tell a threat from an opportunity. It just processes everything the same way. Nobody on your team did anything wrong. Big organisations are built like this on purpose. Whether you fix it is the only question left.
Your organisation isn’t against the new thing. It just can’t tell it apart from a threat.
Your organisation didn’t get big by taking risks. It got big by building excellent systems for saying no: procurement, sign-off, architecture review, the quarterly business review nobody enjoys and nobody skips. Those systems work. They are good at their one job, which is catching expensive mistakes before they happen.
The trouble is they cannot tell an expensive mistake from the one idea in the building actually worth backing. The procurement process that takes six weeks was built to stop someone buying the wrong ERP system. It has no idea the venture needs a cloud account by Tuesday or the whole experiment stalls. The architecture board was built to keep the core platform coherent. It has no idea the thing in front of it will be binned in four months regardless. Neither system asks a single relevant question. They just do their job. On everything. Including the thing you actually wanted to survive.
Nobody sabotaged your programme, and nobody even noticed doing it. It went through the same mincer as everything else, because the mincer doesn’t check what it’s mincing.
The governance reflex
The second anything shows promise, the organisation tries to fold it into normal governance: stage-gates, quarterly reviews, procurement, architecture sign-off. Every one of those exists to protect the core business from risk. Point the same machinery at a venture and you don’t get protection, you get strangulation, because speed was the only thing the venture had going for it.
Sending the B-team
Innovation teams get staffed with whoever the business can spare, and the people who can most easily be spared are, by definition, not the ones you need. Your best people are already busy doing the things that keep the lights on, and nobody wants to pull one off a real job for something the board hasn’t fully committed to. So the venture gets the second string. It fails. Everyone nods sagely and says innovation is hard. It isn’t. You just didn’t send anyone good.
Measuring the wrong thing
Measure a venture on revenue in year one and you get a venture that behaves like a slightly smaller version of the business you already have, because that’s the only way to hit the number. Measurement isn’t neutral. It decides what gets built. Aim it at the wrong target and you’ll hit that target beautifully, and it will be worth nothing.
The sponsor gap
Most enterprise ventures have a sponsor in the way a long-haul flight has a pilot asleep in first class. Someone senior enough to lend the thing credibility, busy enough to think about it once a quarter. The moment the organisation’s defences kick in, and they will, that sponsor isn’t close enough to fight for it. The venture loses every political scrap going, because the one person whose job was to win them is in a different building, doing a different job entirely.
Costly, inconclusive, and it burns the goodwill for whatever you try next
The maths isn’t complicated. A properly boundaried venture costs a fraction of what you will spend keeping a doomed one alive for two years before anyone is brave enough to call it.
>70%
of enterprise innovation programmes never meet what they set out to do
McKinsey
£2.5M
burned on the average failed enterprise innovation programme
Industry estimate
18 months
before everyone quietly knows it’s dead, and finally says so
Observed pattern
This isn’t new. Lockheed worked it out in 1943.
Lockheed built the original Skunk Works in 1943: split from the parent company, governed differently, staffed on its own terms, given explicit permission to ignore the rules everyone else worked under. It produced the U-2, the SR-71, and the F-117. The name, incidentally, came from the smell of the plastics factory next door. The structural principle has aged rather better than the smell.
The Skunkworks Protocol is the same idea, modernised. A hard boundary between the venture and the enterprise so the immune system cannot reach it, and a narrow, controlled pipe that lets budget, brand, and infrastructure through without the governance riding along with them.
The Skunkworks Protocol: Enterprise Edition
The full mechanics, for building something inside a big company that the big company can’t accidentally kill.
A partnership. Nothing that looks like a workshop.
No workshops, no forty-slide deck for the exec committee to nod at and quietly ignore. I run every engagement myself: twenty-five years, four rungs the hard way round, engineer to architect to product lead to exec. The last three specialising in AI, innovation and adoption in enterprise settings, for a client I still work with and do not name, by agreement. On one programme there, the number of passes it took to get from a stated requirement to a signed-off architecture fell by 75%, by their own count.
The opening fortnight is only the first thing you buy. The offer proper is what follows: a working relationship, fractional CPTO, or somebody sat beside the board and whoever on it is carrying the decision. Where you already have a CTO I work to them. Going over a CTO’s head is how you lose the CTO. When a job needs a specialist, I bring one in by name, on my paper, under my insurance. No juniors, no padding, no stranger from the bench turning up in week three.
Structural design sprint
I map how your organisation actually makes decisions. The strategy document describes a different company. Where does authority really sit? Where does information die on its way up? Where exactly will a new venture first meet the immune system? Then I design the boundary, the governance model, and the reporting line that gives it a fighting chance of surviving contact with the rest of the building.
Venture architecture
The venture needs its own operating model, its own success metrics, its own hiring process, and a written relationship with the parent company agreed before anyone is hired. I design that around the actual opportunity, the actual politics you are working inside, and what you are genuinely willing to commit. What sounded good in the pitch is a separate document.
Sponsor coaching
The sponsor decides whether the venture lives more than any other single factor, so that is who I work with, and my job is getting them ready for the political pressure that arrives at entirely predictable moments. When the quarterly review comes and someone asks why it isn’t making money yet, "innovation takes time" is not an answer. I help them build a better one.
PMF review
A structured check against the product-market fit criteria you agreed before launch. Honest reading of the signal, with a written verdict: keep going, change direction, or stop. No sunk-cost reasoning, no dressing it up for the board pack. If the evidence says stop, I say stop. If it says push harder, I say that too. A verdict only means something if it is honest.
“The talent was never the problem. The wiring was.”
If you have run one of these before and watched it die in slow motion, the wiring is almost certainly why. That is the good news. Wiring is the part you can actually fix.