Insights
The delivery problem every services company hits at 30 people
There's a size at which a services company stops working the way it used to, and it takes almost everyone by surprise because nothing obvious has gone wrong.
Roughly thirty people. The number isn't precise — it depends on the work, the margins, how much sits in one founder's head. But the shape is consistent. Below it, quality is held together by a small group who all know how things are done and can see most of what's happening. Above it, that stops being true, and no one decides it has stopped. It just quietly does.
What makes this hard to catch is that the early symptoms look like unrelated small problems. A deliverable goes out with a mistake that would have been caught six months ago. A client is surprised by something they should have been told. Two people give the same client different answers. Each one looks like an individual lapse — someone dropped the ball — and gets treated as such. A word in a meeting, a note to be more careful.
They're not individual lapses. They're the same structural problem wearing different clothes: the informal system that used to hold quality together has quietly exceeded the number of people it can hold.
Why judgement doesn't scale the way you expect
In a small team, quality runs on shared context. Everyone has roughly the same picture of what "good" looks like, what the client cares about, where the traps are. Nobody wrote it down because nobody needed to — you could lean over and ask, or you'd overhear the answer being given to someone else.
That shared context has a hard ceiling, and it's lower than people think. It's not about talent or effort. It's that the number of relationships in a group grows far faster than the group does — the informal "everyone knows what everyone's doing" network that works beautifully at eight people is already straining at twenty and broken at forty. You cannot fix it by hiring better people. Better people in the same broken structure produce the same failures, slightly later.
The reflex that makes it worse
The instinctive response to quality slipping is more oversight. The founder or a senior person starts reviewing more, checking more, sitting in on more. And it works — for a few weeks. Quality recovers because you've temporarily re-inserted the shared judgement by force.
Then it fails harder, for a reason that's almost cruel: you've made the most experienced people the bottleneck for everything, which means you've capped the company's throughput at whatever they can personally review. Growth stops, or quality collapses again the moment they look away, and usually both. You've bought a few good weeks at the cost of the thing you were trying to protect.
The reflex is understandable and it's exactly wrong. The problem isn't insufficient oversight. It's that quality was living in people's heads instead of in the way work is structured.
What actually has to change
The shift is unglamorous and most founders resist it because it feels like bureaucracy — the thing they started a company to escape. But there's a difference between bureaucracy and structure, and it's worth being precise about it. Bureaucracy is process that serves the process. Structure is process that lets judgement scale past the people who originally held it.
Concretely, a few things:
Make the implicit standard explicit — but only where it's failing. You don't need to document everything. You need to document the specific things that are now going wrong because they used to live in shared context. What "done" means for a deliverable. What a client must always be told. What gets checked before anything ships. Not a manual — a short, living list of the things that used to be obvious and no longer are.
Move review from people to points in the flow. Instead of a senior person reviewing whatever crosses their desk, build a small number of defined checkpoints that work passes through regardless of who did it. The reviewer changes from "the most experienced person, whenever they have time" to "whoever owns this checkpoint." That's what unbottlenecks the seniors.
Give quality an owner who isn't the founder. Below thirty, the founder is the quality function. Above it, that role has to belong to someone whose actual job — not side responsibility — is that work holds up as volume rises. This is usually the first real operations hire, and companies that make it late spend the intervening period bleeding the reputation they built in year one.
Why it's worth the discomfort
The companies that navigate this well don't do anything visible. That's the point — the transition is invisible when it works, which is exactly why it's underinvested in. Nobody gets credit for a crisis that didn't happen.
The ones that navigate it badly are recognisable from outside. They're the services company that was great when it was small and got mysteriously worse as it grew, and everyone — including the founder — attributes it to "losing the magic" or "getting too corporate." It wasn't the magic. It was a structural transition that arrived on schedule and got treated as a series of individual mistakes until the reputation was gone.
Thirty people is early enough that you can build the structure before you need it. By the time you obviously need it, you're rebuilding it live while it's on fire, which costs far more and shows.
Entrepreneur, Co-Founder & COO of BoostMySites, associated with BoostMySites Global (Hong Kong). He writes about AI automation, operations, scaling and working with early-stage founders. More about Darshan →