Revenue per employee flatters the P&L and reads to a buyer as risk.

TL;DR: Low headcount has become a scoreboard, and some founders are now chasing it as preparation for a sale inside two years, which is where it can quietly cost them. Revenue per employee reads well on a one-pager and says nothing about replacement cost, whole departments resting on one person, or an efficiency layer built from tooling that can be worthless next quarter. There is a real exception, when the acquirer was going to replace your team anyway. Outside it, the question a buyer is asking is whether people who are not you can run the business.

Richard Lindner raised something on the podcast a while back that stuck with me. There is a trend among entrepreneurs of bragging about running a business with an extraordinarily lean team. The pendulum has swung, he said, and it is now a badge of honor to operate with as few people as possible.

His question was the right one. Is there such a thing as too lean, and what is the trade-off between operational efficiency and building something that carries transferable value long term.

Let me be careful, because the answer is not that lean is wrong. It depends entirely on which of three companies we are talking about. A while ago we started calling what was happening the great flattening, businesses collapsing into one-person departments, and AI has enabled a great deal more of it, though it was already underway before AI.

What has gone wrong is narrower than the trend. Richard is hearing from operators who are cutting toward leanness as preparation for a sale, and that is where it can quietly cost them.

None of this can be answered until you know which of three companies you are running, because they get priced by different logic.

The third is the best place to be, and the least celebrated, because it does not photograph well.

For the first, the valuation math is not subtle. Market multiples are significantly lower for owner-operated businesses than for professionally managed ones, because a professionally managed business can be bought as an investment and an owner-operated business is bought as a job. There are far more investors in the market than people shopping for a job to buy. I have written before about what professionally managed actually means to a buyer, and it is closer to decision infrastructure than to good delegation. If you are in that first bucket, leanness is not your problem. Being the business is.

The second bucket is the interesting one, because on paper it looks like the winner.

Revenue per employee at a tech-pancaked company looks superb, and it speaks well toward a high valuation. It is the number founders put on the one-pager, and they are not wrong to.

Here is what I look at next, and what I believe most acquirers do. Replacement cost. Dependencies on individuals for entire departments. What happens if there is a tech interruption, and there are tech threats all the time. And how committed you are, capital-wise, to the tooling holding it all up.

That last one gets less attention than it deserves. Tech is becoming more expensive as a budget item, almost like a capital expenditure, except it is not a fixed thing you can turn around and sell, because it is probably worth nothing, having gone obsolete last week.

Richard put the sharpest version of the problem. Technology breaks, so it needs to be there to amplify. Some roles it can completely own, others it should only amplify. Departments of one are fine. Departments of one AI employee are not. His reasoning: a human employee becomes obsolete over a period of years, and only if they take no interest in their own development, while an AI employee can become obsolete in a matter of weeks, because the tools change every six minutes, every six days, every six months.

The case Richard worries about most is specific. Non-lifestyle businesses cutting to three or four human employees, with a stated goal of exiting inside eighteen to twenty-four months. His read is that it will be seen as a massive devaluation.

I came across a company that saved forty million dollars a year by replacing most of its front-end customer service people with AI. Then everything fell apart. Terrible reviews, terrible customer experiences, lost customers, and eventually they went back to humans. I could not tell you the name of it, and it is not the point. To be lean in customer support is usually to give poor customer support, whether that comes from AI, offshore contractors, or simply not having support at all, the way plenty of tech companies do not.

There is a genuine exception, and leaving it out would mislead you.

Extra lean fits well into an acquirer who was going to fire all the employees anyway. When we sold Traffic and Conversion Summit, the buyer's position was essentially that they did not need us except perhaps to speak, and did not need any of our employees or contractors either. They had a team already, and they dropped it into their company.

So ask directly about your own likely buyer. Would they want your people, or would they simply run it with the team they already have? If you can identify that your potential acquirer probably does not want your people anyway, then extra lean is probably okay. Richard pushed it one step further: a buyer may decide your team has to unlearn what you taught them and form new habits, which makes its own case for an extra layer of lean.

But notice what that exception is not. It is not general permission. It is a fact about one specific buyer, which makes it a bet on which buyer shows up, and the reason to get ready early is that you do not get to choose. Richard's caveat is the one I would hold onto. Unless the acquirer is completely AI-augmented themselves, they still want to know humans can run the company. They want to see SOPs, an operating system real people can run, because that is the transferable value. Prompts and scripts and things running in the background that not everyone can decouple are not.

And if you are building something to sell, as Richard put it, what you are selling is the ability to execute. Leanness describes how you happen to execute today. It is not the thing itself.

For years the work I have pushed founders toward is making themselves unnecessary, and I still believe in it. But there is a version of it that misses the point: making yourself unnecessary by making a piece of software necessary instead. That is not independence. It is the same dependency in a cheaper coat, and it gets repriced the same way.

The version worth having is the one nobody posts about. It is unglamorous, and the only one of the three that leaves you a choice about what happens next. Optionality is not a low employee count. It is a business that keeps running regardless of who, or what, stops showing up.

— Roland

Don't be the cheap one.

The roll-ups are buying old-school businesses below market, adding AI, and marking them up overnight.

If that's the knock on your door, you have two ways out.

Make the upgrades yourself before you sell.

Or keep a real piece and let the buyer's work multiply it. Sell 60 or 70 percent, keep the rest, and if an experienced team six-times the business, the piece you kept is worth more than everything you sold.

Want the full breakdown? Watch the reel.