How to price a business you never intend to sell

Take away the exit and most of conventional deal pricing stops working, because most of it is a forecast about what somebody else will pay you later. A permanent holder has no later. The price has to be justified entirely by cash the business produces while you own it, which is a harder test and a considerably more honest one.

What a normal model is actually doing

Underwrite an acquisition the standard way and the structure is always the same. You project earnings forward five years, assume an exit at some multiple, discount it back, and see whether the return clears your target. It is a sound method and it is used because it works.

Notice, though, where the money comes from in that model. A large share of the projected return, frequently the majority, is the terminal value: the assumed sale at the end. And the assumed sale rests on two things you do not control, namely what multiples look like in five years and whether anybody wants this type of business then.

The multiple expansion component deserves particular scepticism, because it is where a lot of underwriting quietly does its work. Buy at five and sell at seven and you have made money without the business improving at all. It happens, and it is also a bet on market conditions dressed up as a plan.

The test that replaces it

Strip out the terminal value and the question becomes concrete: over how many years does this business pay for itself out of what it actually produces, and can it comfortably service whatever I used to buy it while doing so.

Two disciplines make these useful rather than decorative. Use a multi-year average rather than the trailing twelve months, because the trailing twelve is why the seller chose this moment to sell. And run coverage on the average rather than the good year, because the whole point of the exercise is surviving the bad one.

What falls out is a hard ceiling on what you can pay. Not a valuation opinion, a ceiling, below which the deal works and above which it depends on things you have decided not to rely on.

And the ceiling states itself, because of something the payback test hides in plain sight. If flat cash pays the price back over a number of years, then that number of years is the multiple you paid. Pay five times rebuilt cash and the payback is five years. Which means the payback period, the purchase multiple, and the unlevered annual return are one number wearing three costumes:

Which is why the discipline is easier to hold than it sounds. You do not have to argue about valuation methodology at the table. You have to answer one question in advance, how many years you are willing to be wrong for, and everything else is division.

Put the other posts on the same axis and they stop looking like separate arguments. Constellation's 30 percent hurdle for a small acquisition is a 3.3 year payback. The median search fund deal at 6.2 times is a 6.2 year payback, or a 16 percent return, on a business bought by somebody who intends to sell it in five. The gap between those two is not a disagreement about value. It is that one of them is counting the exit and the other is not.

That ceiling is frequently lower than what the market will pay, which brings us to the awkward part.

You will lose deals, and that is the mechanism working

A buyer underwriting to an exit can justify a higher price than one underwriting to cash, because they are counting money you have refused to count. On any given competitive process, they will win.

This is the part that requires actual conviction rather than agreement in principle. Watching a business you wanted go to somebody paying two turns more, repeatedly, is not comfortable, and the temptation to adjust the method after the third loss is strong. But a method that moves when you lose is not a method, it is a preference you abandon under pressure, and the hurdle you lower once stays lowered.

The deals you lose to somebody underwriting an exit are not deals you lost. They are deals you declined to make on terms that only work if they can sell it.

What makes the discipline survivable is that the two of you are not competing for the same businesses as often as it appears. A sponsor needs a business that will be attractive to a future buyer, which biases them toward growth, scale and a good story. A permanent holder needs one that will produce cash reliably for a very long time, which is a different and much duller test. Plenty of businesses pass the second and fail the first, and there is far less competition for those.

What you can pay more for

The framing so far is one-directional, and it is not the whole picture. There are things a permanent holder can pay up for that a sponsor genuinely cannot, and being clear about them is what stops this from being a strategy of only ever bidding low.

1. Improvements with a long fuse

   A change that takes three years to show up in earnings is worth almost nothing on a five year clock and everything on no clock. Anything requiring patience is systematically underpriced by buyers who need to show a result before they raise again.

2. Durability over growth

   A business growing at two percent a year forever is a poor story and an excellent asset. Nobody is bidding against you for it, because it will not look good in a sale document five years from now.

3. What the seller wants that is not cash

   Continuity, the name kept, the staff looked after, a handover on their terms. A holder can promise those credibly because they intend to be there. That is frequently worth more to a founder than the last few percent on price, and it costs nothing.

4. Structures that need a long counterparty

   A payment stream running for decades only makes sense from a buyer who will still exist and still own the business. That opens deals that a fund cannot practically do, and it opens them at prices set by a much smaller field.

The third and fourth are where the real advantage lives, and both are unavailable to a bidder with a fund life. That is the compensation for the lower ceiling: you are not bidding lower into the same auction, you are bidding into a different and less crowded one. The wider version of that argument is in permanent capital versus the fund clock.

The divisor, and why it has a floor

Reduced to a rule, pricing to keep sets a maximum multiple of rebuilt earnings and refuses to go above it, with the maximum derived from payback rather than from what comparable deals cleared.

The floor under that divisor is the part people try to negotiate away, usually with a good argument about a particularly high-quality business. The argument is often correct about the business and still wrong about the price, because the floor is not a judgement on quality. It is the point below which an ordinary bad year stops being survivable, and quality does not prevent bad years, it just makes them less frequent.

Set it in advance, in writing, when nothing is on the table. Written afterwards it is not a floor, it is a description of the deal you had already decided to do. That is the same discipline as the binary gates that run before any valuation work at all, which are in the seven things that end a deal on the first call, and it fails the same way when it bends.

None of this makes pricing easy or removes judgement. What it does is move the judgement to a place where you can defend it: what the business produced, averaged over years, verified rather than projected. Compared with an assumption about a market five years out, that is a much smaller thing to be wrong about. How to get the earnings figure honest in the first place is a range rather than a number, and what the market is actually paying is in what a small business actually sells for. What I will pay and will not is on the buy page.