The Speed You Had at the Start Is Still Available
At some point the thing that made your company good turned round and started working against new ideas. The processes that brought order also brought inertia. The layers that let you scale also let decisions get lost inside them. The speed hasn’t gone anywhere. It’s just buried under a structure built to stop mistakes. Movement was never on the brief.
You're moving slower than you should be
The drift never feels dramatic while it’s happening. Ship time creeps from days to weeks, then weeks to months, and a feature that would have taken a week three years back now eats a quarter, and nobody in the building can quite tell you why.
The people haven’t changed. The ambition hasn’t changed. The skills, if anything, are better than they were. What’s gone missing sits somewhere between the garage energy you started with and the org chart you’ve got now: the ability to move fast on something new, without anyone noticing it slip away.
Call it a wiring problem, because that is what it is. The people are fine. It happens to every company that grows past a certain size without somebody deliberately guarding the conditions that made it sharp in the first place.
Growth and speed work against each other by default
When there were fifteen of you, the loop was tight enough to trust. A customer moaned and the person who could fix it heard about it by teatime. A competitor twitched and the whole office knew by breakfast. Decisions got made in the corridor, over a bad coffee, at pace.
Growth stretched that loop out, hire by hire. Every new layer added a bit of lag. Every process, sensible on its own, added a bit of friction. None of it was wrong individually, you needed the structure to get bigger. But add it up and the loop that used to make you fast now runs so long that by the time a customer’s signal reaches someone who can act on it, it is three weeks old and wearing a slide deck.
The company that used to answer a market shift inside a week now needs a quarter. The one that could test a new idea in a month now wants a year for the same job. The wiring changed under you, and nobody clocked it happening because every single change, on its own, looked like good sense.
A small separate team. Real boundary. Startup speed.
The Skunkworks approach draws a proper line inside your company and defends it. A small team, three to seven people usually, runs with the autonomy of a startup against one specific bet. Their own budget. Their own authority. Their own definition of success, agreed before anyone starts.
The boundary has to be real, not a name on an org chart, and not a “fast-track team” or an “innovation pod,” whatever that’s supposed to mean. It is a genuine wall that keeps the governance, the process, and the internal politics of the parent company away from the venture. The team talks to customers directly, ships without a chain of approvals to climb, and is allowed to fail fast and cheap, which beats the slow expensive version every time.
The parent company brings the money, the brand, and the plumbing. The venture brings speed and a willingness to be wrong in public. The boundary is the only thing holding that trade together. Take it away and the parent company wins, because it always does. Not a criticism. Bigger organisations are built to protect what already works, and a fragile new idea is exactly what that machinery is designed to crush.
Three stages with clear checkpoints
Phase 1: The Contained Experiment
Months 1–6Small team, a budget nobody can touch, one clear guess worth testing. They borrow the parent company’s plumbing but own every decision themselves. Success here is measured in what you’ve learned. The invoices can wait. The only governance is a monthly catch-up with the sponsor.
Phase 2: Scaled Proof
Months 6–12The guess has legs. Now find out whether it holds up against real customers, at a volume worth caring about. The team grows a little, but the boundary doesn’t move. This is the exact point where most SMEs panic and fold the venture back into the core business. Don’t. That’s how you kill it.
Phase 3: Integration or Independence
Month 12+Either the venture pays for itself and becomes a proper business unit, or it doesn’t, and what you’ve learned gets absorbed back in while the team moves on. The outcome worth actively avoiding is the middle ground: a permanent "innovation team" that produces reports instead of revenue.
Your instincts as a founder will make this fail
Everyone skips this bit. It is also the bit that decides whether the venture works, or quietly burns your budget until somebody writes a post-mortem nobody reads.
Your closeness kills speed
The urge to stay involved, to review, to nudge things back on course, is the same urge that throttled your last three attempts at this. Founders who can’t step back from the new venture almost always end up strangling it, out of love.
Your best people are the wrong people
The operator running your core business was built for a different game entirely. A Skunkworks team needs people who can live with not knowing, who don’t mind being wrong in public, and who’ll ship something before it’s ready. Putting your best operator on it is like sending the accountant up in a stunt plane.
Wrong metrics, wrong behaviour
Measure the new venture on the same numbers as the core business and you get core business behaviour, guaranteed. Set a revenue target for month three and you’ve told the team to chase a quick win instead of testing anything. Measure what they’ve learned. Worry about the money later.
Twelve months minimum
A quarter won’t tell you anything useful. Neither will six months. Twelve months of properly protected time is the minimum before you’ve got a fair verdict, and if you can’t commit to that, don’t start at all. A half-hearted attempt burns more money and more morale than never trying, and everyone in the building remembers which one you did.
Five prerequisites that are non-negotiable
Ring-fenced budget
Not "we’ll find the money somewhere." An actual number, signed off in advance, that nobody can raid the moment the core business has a wobbly quarter. The instant the budget becomes a conversation, the venture is already dead, it just hasn’t been told yet.
Written mandate
One page. What the team can do, what it can’t, and what success looks like at six months and twelve. Skip this and every decision becomes a political negotiation, which is a slow way to run a team you’re paying to be fast.
Pre-agreed success criteria
Agree, before you start, what "working" actually looks like. Not revenue, not a user count plucked from the air, but one specific guess and the evidence that would prove or kill it. If you can’t nail that down before day one, you’re not ready, whatever the calendar says.
Hiring independence
The team hires its own people. Not HR, not the existing management chain handing down whoever’s spare. The people running the venture choose who’s in it. This is not up for discussion.
One protector
One senior person, and one only, whose entire job is keeping the rest of the business off the venture’s back. Not a committee, and not a steering group either, which is just a committee that meets more often. One person with enough clout to say "leave them alone" and make it stick.
What doing nothing actually costs
A clean failure runs you somewhere between £150k and £500k. Sounds grim, but a clean failure is genuinely the best outcome a bad bet can produce. You learn something real, you redeploy the team, and you walk away with knowledge that cost you but is now yours.
An inconclusive result costs two to three times as much, and it is the outcome everyone actually fears. The venture that will not quite die and will not quite work, dribbling away budget and attention for eighteen months until somebody finally finds the nerve to pull the plug. The money is not even the real cost. It is the attention and the goodwill it eats on the way down.
The opportunity cost is the one nobody can put a number on, which is exactly why it gets ignored. The market that moved while you were still deliberating. The competitor who shipped while you were still planning to plan. The good people who left because they wanted to build something, and you kept asking them for a deck about building something instead.
£150k–£500k
Clean failure
2–3x more
Inconclusive result
Incalculable
Opportunity cost
What we actually do
Structural Diagnosis
We map how decisions actually move through your business. Not the org chart, the real thing, the one nobody’s drawn because it’s slightly embarrassing. Where the information gets stuck, where the authority actually sits, where the gap between knowing and doing has widened past useful.
Venture Architecture
We design the boundary itself: the shape of the team, the budget model, how light the governance can be and still count as governance, the reporting line, the success criteria. The blueprint that gives the venture a fighting chance inside your company. Quiet reabsorption is what happens to the ones without one.
Founder Coaching
The hardest part of running a Skunkworks team inside an SME is managing yourself. The team is the easy bit. We coach the founder or sponsor on an ongoing basis, aimed squarely at the moments your own instincts are about to undo the whole thing.
Milestone Review
A proper review against the criteria you agreed at the start. An honest read on where things stand, with a clear recommendation: carry on, change tack, or stop. No sunk-cost thinking, no face-saving. Just what the evidence says.
Self-Diagnosis
Seven questions. The last one is the hardest.
Can you actually name the last genuinely new thing your company shipped? A tweak to something that already existed doesn’t count.
Question 1 of 7
“The speed never left. It just needs somewhere to go.”