There is a particular conversation I have with founders growing at thirty per cent or more, and it always starts the same way. Something has gone badly wrong in an area that used to be the strongest part of the business, and nobody can work out why.
The delivery team that everyone praised is missing dates. The support function that customers used to comment on unprompted has become a source of complaints. The two people who always knew what was happening no longer do.
The founder's instinct is that something has degraded. Standards have slipped, or the wrong people were hired, or somebody stopped caring. That is almost never what happened. What happened is that growth applied load to a structure, and the structure failed where the load was heaviest, which is the part that was carrying most of the business.
Good things break first. That is not a paradox, it is mechanics.
01Why does the strongest part of the business fail first?
Because it was strong, so you leaned on it.
At fifteen people, an informal high-trust operation is genuinely better than a documented one. Decisions are quick, exceptions are handled sensibly, and nobody wastes time on process for its own sake. It is not a shortcut, it is the correct design for that size. So the business routes more and more through it, because it works.
Then volume doubles. The informal operation now requires the same few people to hold twice as much context, handle twice as many exceptions, and be available twice as often. It does not degrade gracefully. It works, works, works, then stops, usually within a single quarter, and from the inside it looks like a sudden collapse in standards.
This is why "we need to hire better people" is such a common and such a wrong conclusion. The people are the same people. The load changed.
02What is the order things break in?
Consistent enough across businesses to be worth planning around.
First, decision-making. Typically somewhere between twenty and thirty-five people. The informal way calls got made depended on a handful of people having the full picture, and the picture is now too big. The symptom is that everything takes longer while nobody is working slower. Things wait.
Second, handovers. Between functions, and between people within a function. At small scale, handovers work because both parties know each other and fill the gaps by talking. At scale, the gaps stop being filled, and what surfaces is that the handover was never actually defined, only performed. This is where quality problems appear and where customers notice.
Third, cash. Growth consumes working capital, always, and it does so quietly. You are paying for delivery capacity in advance of the revenue it produces, funding a larger debtor book, and carrying more overhead against a margin that is being squeezed by the first two failures. Businesses at thirty per cent growth get into cash difficulty while being genuinely profitable, and it takes founders by surprise every time.
Fourth, culture. Last, not first, though it is what founders worry about earliest. Culture holds up remarkably well through the first three failures. What kills it is not headcount, it is watching the first three go unaddressed. People do not leave because the business got bigger. They leave because it got worse and nobody seemed to notice.
Culture is the fourth thing to break, and by the time it does, it is the symptom of three problems you have already been told about.
03How do you find the load path before it fails?
Trace one order, end to end, and count the human dependencies.
Take a single typical piece of work, from first enquiry to cash collected. Write down every point at which something has to move from one person to another, and for each one, answer two questions. What has to be true for this to work? And what happens when the person who normally does it is not there?
You are looking for the points where the honest answer is "Sarah knows". Every one of those is load bearing, and every one is a point of failure at twice the volume. In most businesses under £10m, there are between four and seven of them, and the founder can usually name three off the top of their head. It is the other three or four that cause the trouble.
Do this exercise annually and the results are strikingly consistent: the load path does not move much, but the amount of load on it doubles, and the number of dependencies you can tolerate falls as speed becomes more important.
04Why is reinforcing better than replacing?
Because during a growth phase, replacing the structure means removing the thing currently holding the business up and substituting something untested, while the load continues to rise.
I have watched businesses respond to the first breakage by commissioning a full process redesign, or a new operating model, or a system implementation covering everything. Eighteen months later they have a documented operation nobody follows, and they lost eighteen months of growth to it. The intent was right. The scope was wrong.
Reinforcement looks different. You take the single point where load actually falls, and you strengthen it specifically. If quoting is the constraint, you write down how quoting decisions get made and hand that to one more person, which takes a fortnight. If handover from sales to delivery is where quality is lost, you define that one handover, not all of them.
The test for whether you are reinforcing or replacing: can you point at the specific failure this work addresses, and would you know within six weeks whether it worked? If not, you are rebuilding, and you should be doing it after this growth phase rather than during it.
05What about the parts you should let break?
Some things should be allowed to fail, and founders are reluctant to hear it.
There are ways of working that were correct at your old size and are simply not worth preserving. The all-hands meeting where every decision was discussed. The founder personally welcoming every new customer. The shared inbox everybody watched. These are frequently the things people are most attached to, because they are what the business felt like when it was fun.
Trying to preserve them at three times the size produces exhausting theatre. The all-hands becomes a two-hour broadcast, the personal welcome becomes a template, the shared inbox becomes a thing nobody watches and everybody assumes somebody else is watching.
Better to name what the practice was actually for and rebuild that intent in a form that suits the current size. The all-hands existed so people understood why decisions were made, which at eighty people is a written decision log, not a longer meeting. This is harder than it sounds because it means saying out loud that something people loved is over, but pretending otherwise is worse.
06The version of this that matters
Growth is the only condition under which all of this is worth doing. A flat business can run on an inadequate structure for a decade.
Which means the failures are not a sign that something has gone wrong. They are a sign that something is going right, arriving on a schedule you can predict. Founders who understand that spend their energy on the next thing due to break rather than on working out who is to blame for the last one.
The order does not change. Decisions, handovers, cash, culture. If you are at thirty per cent growth and decision-making has just started sticking, you now know what the following four quarters look like, and you can be in front of it rather than behind.