The founder is the bottleneck

Early on, being in everything is correct. You are the fastest decision-maker in the building because you hold the most context, and a small company with no process runs on exactly that. Every founder who got to a million dollars did it partly this way.

The trap is that nothing announces when it stops working. There is no day the advantage turns into a constraint. The company just gets slightly slower each quarter while you get busier, and both of those feel like growth until you look at what is actually happening: work is waiting.

Forget how you feel. Feelings measure your workload, not the company's throughput. Measure waiting instead.

For one week, write down every decision that came to you and how long it had been sitting before it reached you. Not the time you took — the time it waited. Then count how many of those decisions someone else could have made if they had known the rule you were applying.

That second number is the honest measure of the bottleneck. In most companies at this size, it is somewhere north of half, and the founder is genuinely surprised.

The obvious cost is speed. The larger cost is that your team stops developing judgment. When every meaningful call routes through one person, nobody else practises making them, so the day you try to hand something over, you hand it to someone who has had no reps — and it goes badly, and that failure gets read as proof that you had to keep doing it yourself.

This is well-trodden ground outside startups. MIT Sloan's Elsbeth Johnson, in a Harvard Business Review conversation on delegation, makes the point that leaders under-delegate not because they lack time but because handing over decisions feels riskier than handing over tasks. The startup version has an extra edge: your identity is fused to the company, so letting go of a decision feels like letting go of the thing you built.

This is the order I use with founders, and the order is not decorative. Most people jump to step two, hire a strong person against an imagined job description, and are disappointed within six months.

Two weeks, every hour, roughly. Not the job you imagine you do — the one you actually did. Everything that follows is decided by that log, not by your instincts about it.

Sort it afterwards into three buckets: work only you can do, work you are doing because you are fastest, and work you are doing because nobody ever took it. The middle bucket is the biggest and the most painful, because being fastest is a genuinely good reason and it is still the thing holding the company at its current size.

One owner per function — sales, delivery, money, product. One name each, written down, even if two of them are still you for now. Shared ownership is how things quietly become yours again.

Then one number per owner, reported weekly in a single line without being asked. If you cannot name the number, you have not defined a job, you have defined a helper. That distinction is most of the difference between a team that absorbs decisions and a team that forwards them.

Write down what each owner decides alone, what they tell you about afterwards, and what genuinely needs you first. Put a dollar figure on it.

Almost every founder who does this discovers the real threshold is roughly ten times higher than how they have been behaving. They have been approving four-figure decisions while telling themselves they only get involved in the big ones.

Leave for two weeks and do not check in. This is the test, not the reward. What breaks is your roadmap for the next quarter; what holds is the company you have actually built.

Then fix what broke and leave for a month. One round of this is a holiday. Two rounds is a business that has stopped depending on you — and, not incidentally, a business worth materially more if you ever sell it, because founder dependence is one of the first things a buyer prices in.

You answer the question anyway. Someone brings you something that has an owner, and answering takes eleven seconds. Answering once is kindness. Answering twice rebuilds the bottleneck you just dismantled. Send it back with the name.

The other failure is hiring your way out before step one. A senior hire dropped into an undefined role becomes another person waiting on you, at a higher salary.

An inventor I know patented a medical device, took the company public, and stepped back young. On an ordinary afternoon a stray golf ball took the sight in one of his eyes. While he recovered, his wife ran the company. She had never asked for the job, never wanted the title, and did it well.

That is the version of this nobody plans for. The question is not whether you want to step back — it is whether the company survives a week you did not choose. Most founders I work with discover the answer at the worst possible time.

Everything below is a way of finding out on purpose, while the stakes are a delayed launch rather than a business that stops.

Start the log tomorrow. Not a system, not a new tool — a note on your phone, every hour, for two weeks. Everything above depends on knowing what you actually do, and no founder I have worked with has guessed it correctly in advance, including the ones who were sure they would.