You're faster than they are. That won't last by accident.
The speed that makes you dangerous isn't a personality trait. It's structural: how fast a customer complaint turns into a decision. It doesn't survive growth by accident. Every startup that scaled cleanly protected that loop on purpose. Every one that didn't rebuilt it later, at far greater cost, usually after losing an argument to a competitor with four people and no patience.
The loop is the asset
It isn't the product, the team, or the funding round. It's the loop.
At your stage, a customer moans and the person who can actually fix it hears about it by teatime. A rival launches something and the whole team knows before lunch. An engineer has a decent idea on Monday and it's live by Thursday.
That loop, how fast information gets from the outside world into a decision and back out again, is worth more than your technology stack and more than your last funding round put together. And it leaks a little every time you hire someone, add a layer of management, or let a process take root that never used to be there.
Protect the loop or lose the advantage. Nobody automates this bit for you.
Someone is coming for you the same way you came for the incumbent
Somewhere right now, four people are staring at your market with exactly the impatience you had eighteen months ago. They're faster than you were, because the tools got better while you weren't looking. Cheaper than you were, because a server costs a fraction of what it used to. And considerably angrier than you are, because from where they're sitting your mistakes are obvious.
The only thing standing between you and being somebody's cautionary tale is how fast your loop still moves. Keep it tight and you'll see them coming, and deal with it before they get traction. Let it stretch and you'll hear about them from a board member who read it on a plane.
The three stages nobody warns you about
0–15 people
The loop is tight
Everyone talks to customers because there's nobody stopping them. Decisions happen over lunch, and nobody has to book a room for them. The feedback cycle runs in hours, sometimes less. This is the whole asset. Everything you build from here either keeps it alive or slowly strangles it.
15–50 people
Layers are forming
The first manager turns up. Conversations that used to happen in the corridor now need a channel and an agenda. Some decisions start requiring a slide, which should worry you. The loop has stretched. It hasn't broken yet.
50+ people
The structure has changed
The person who hears the complaint and the person who decides what to do about it are now separated by two layers of management and a shared calendar. The loop is broken, and it stays broken unless somebody fixes it on purpose.
A small separate team with a real boundary
The Skunkworks Protocol isn't a process, a framework, or anything you'd find in a methodology binder. It's a structural decision: you build a small team and put a real wall between it and the rest of the company. That wall is what keeps the speed alive.
The team moves at startup pace because, structurally, it still is one: small, its own boss, talking to customers directly, and free of the governance that keeps the core business safe while making anything new almost impossible to ship.
The boundary can't just be a mood. "We trust them to move fast" is a sentiment, not a boundary, and sentiments don't survive contact with a quarterly review. What survives is structure: separate budget, separate authority, separate scoreboard.
Three stages from borrowed to independent
Phase 1: Borrow Everything
Use the parent company's finance, legal, brand, plumbing, whatever's already lying around. The skunkworks team spends zero hours on anything that isn't the actual bet. Run lean, move fast, find out if the idea has legs before anyone spends real money finding out the hard way.
Phase 2: Build What You Know
Once the signal is real, start building the functions the new venture actually needs. The old company's versions were built for the old company. Hire for where you're going. The people who got you this far may not be the right people for what's next, and that's nobody's fault.
Phase 3: Stand Alone
The venture runs its own numbers, has its own team, makes its own calls. It earns its independence or it gets folded back in. There is no comfortable middle where it limps along half-owned. Staying vague at this stage is how good ideas quietly die.
Eight signs the inflection has started
If three or more of these are true, the loop is already degrading.
You can't remember the last time you actually spoke to a customer, and it's been more than a fortnight.
Decisions that used to happen over coffee now need a meeting booked three days out.
A feature has been "in discussion" for three months and nobody can say why it isn't live.
A sharp new hire mentioned the pace surprised them. They meant it wasn't fast enough.
Somebody proposed a "process" for something that used to just happen because a person cared.
Your roadmap reads like a peace treaty: everyone's interests represented, nobody quite happy.
Engineers are asking permission more often than they're building.
Your last three product calls went to whoever talked loudest in the room. The person closest to the customer was not in it.
Four deliberate acts of structural preservation
Name the loop explicitly
If you can't explain, in one sentence, how a customer complaint turns into a decision, the loop is already going soft. Write the sentence down. Pin it up somewhere. Guard it like it's worth money, because it is.
Decide what won't change
Growth wants to change everything, given the chance. Don't let it near a few things: how close the team stays to customers, how fast a decision gets made, who is actually allowed to ship. Decide these before scale forces a worse answer on you.
New markets get separate teams
The quickest way to kill a new idea is to bolt it onto the team already running the old one. New bets need their own people, their own budget, their own measure of success. Otherwise the core business's gravity wins every time, because it always does.
Track signal-to-decision lag
Measure the gap between "the customer said something" and "somebody did something about it." If that number is growing, you are losing the exact thing that made you dangerous. It is the metric nobody is watching, and it is the only one that matters.
What we actually do
Structural Audit
We map how the decision loop actually works today, not how the org chart says it works. Where information comes in, where it gets stuck, who really holds the authority, and how wide the gap has grown between a customer saying something and someone acting on it. No surveys, no workshops. Watching and asking.
Second Bet Design
If you're ready for a new venture or market, we design the actual boundary: team shape, ring-fenced budget, how it's governed, what counts as success. The structure that gives the new thing a fair fight instead of a slow death by committee.
Founding Team Advice
Hiring for a skunkworks isn't the same job as hiring for the core business. We advise on team shape, seniority mix, and the kind of person who should run it. Get this one wrong and it's the most expensive mistake on the list.
Ongoing Structural Review
The loop doesn't collapse, it erodes, a little at a time, and you won't notice until it's too late unless somebody's actually watching. Quarterly check-ins against the original design, honest about what's slipping, with a plan to correct it.
Self-Diagnosis
Six questions. Answer them straight. There's no prize for flattering yourself.
Can you name the actual customer problem your next bet solves, without reaching for the words "platform," "ecosystem," or "solution"?
Question 1 of 6
“The loop that made you dangerous was always structural. Guard it on purpose, or watch it vanish right on schedule.”