Where the cash goes after a good year

What you do with the money at the end of a good year is the single decision that compounds, and in most privately held companies it is not really a decision at all. The cash accumulates, some of it gets taken out, some of it buys a piece of equipment somebody asked for, and the rest sits in the account. Done deliberately instead, that same choice is what separates a company that grows from a group of companies that compound.

There are only five places it can go

Strip away the language and every owner has the same short menu. Money coming out of a business can be put back into the existing operation, spent buying another business, used to pay down borrowing, distributed to the owners, or left in the bank. That is the whole list, and every one of those options has a return attached to it whether or not anybody has calculated it.

The reason this matters more than almost anything else an owner does is that operational decisions are annual and allocation decisions are permanent. Improve your scheduling and you get a better year. Put a year's profit into the wrong thing and you have converted a year of hard work into an asset that will keep producing a mediocre return for a decade.

The default is worse than any of the five options chosen on purpose. Cash left in the account earns close to nothing while inflation works on it, and it does not stay neutral either, because a large balance eventually attracts a purchase that would not have survived a proper comparison. Idle money finds a use, and the use it finds is rarely the best one available.

The hurdle, and why it has to be uncomfortable

The tool for making the choice is a hurdle rate: a minimum return below which you simply do not deploy. The most instructive public example is Constellation Software, which has spent three decades buying small software companies and is unusually open about how it decides.

Their hurdle is tiered by size. Roughly 30 percent for the smallest businesses, those under a million in revenue, around 25 percent for mid-sized ones, and about 20 percent for deals above four million. There is a lower bar, around 15 percent, for very large deals above fifty million, and they might do one of those in a year.

The tiering is the interesting part and the logic runs opposite to intuition. A small acquisition costs almost as much management attention as a large one, so it has to clear a higher bar to be worth the attention it consumes. Size does not earn a discount because it is safer. It earns one because it is more efficient per unit of scarce attention.

Hurdle rates are magnetic. Lower the bar for one deal and the expected return on every deal you look at afterwards drifts down with it.

That observation, which comes from Mark Leonard, is the most useful thing in the whole subject. The intuitive model of a hurdle is a filter: drop it slightly and you let in a few more deals at slightly lower returns. That is not what happens. Once the bar moves, the entire pipeline recalibrates around the new number, the deals people bring you change, the prices you are willing to discuss change, and the average outcome of everything you do afterwards falls. You do not get the old portfolio plus some marginal additions. You get a worse portfolio.

Note what a hurdle is not. It is not a forecast and it is not a promise. It is a rule you set before you have a specific opportunity in front of you, precisely because you will not be objective once you do. Written down in advance, it is a constraint. Chosen in the moment, it is a justification.

A hurdle rate is a price, stated backwards

The reason hurdle rates feel abstract is that they are quoted as returns while every actual negotiation happens in multiples. They are the same number. If a business earns the same amount each year and you pay a multiple of those earnings, your return is simply one divided by the multiple. So a hurdle is a ceiling on price, and translating it makes the discipline concrete:

Now put that beside what small companies actually change hands for. The median search fund acquisition, on Stanford's forty-year dataset, went at 6.2 times earnings. That price carries an unlevered return of about 16 percent.

That single comparison explains more about the quarter of deals that lose money than any amount of discussion about diligence quality. At 6.2 times, the return depends entirely on making the business better after you own it, because the price itself has already spent the margin for error. Constellation's answer is not that they are better operators. It is that they refuse to start from there.

Two honest qualifications, because the arithmetic is cleaner than reality. It assumes flat earnings, so a business that genuinely grows justifies a higher multiple and a lower starting yield. And it is unlevered: debt raises the return on your own money and raises the consequences of a bad year in exactly the same proportion. Neither changes the shape of the comparison, and both are the arguments a seller will make.

What this looks like below the billion-dollar level

None of this requires scale. An owner of one profitable company faces exactly the same five doors, and the discipline transfers directly.

1. Reinvest in what you already own

   Usually the highest return available and almost always underused. Fixing something that leaks in a business you already control has no acquisition premium, no integration risk, and no new relationships to build. If enquiries are going unanswered, that is the best-returning use of money you have, and it is cheap.

2. Buy another business

   The right call once the first one is genuinely running without you, and a serious mistake before that. Buying a second company while the first still needs you daily does not double the income, it halves the attention available to both.

3. Pay down borrowing

   An unglamorous, guaranteed, risk-free return equal to your interest rate. It is the honest benchmark every other option has to beat, and it is the correct answer far more often than anyone finds interesting.

4. Take it out

   Legitimate, and the reason the whole thing exists. The mistake is treating it as the residual rather than a deliberate number: decide what comes out, then allocate what remains, rather than allocating what happens to be left.

5. Hold it, on purpose, for a named reason

   Cash held for a specific known event is a position. Cash held because no decision was made is a slow loss with an eventual bad purchase attached to it.

The first and third options being so strong is the part most people resist, because neither feels like progress. Buying something is visible and satisfying. Removing the reason your quotes go out four days late is invisible and produces a better return, which is an uncomfortable pairing but not a complicated one.

The second half nobody copies

Groups that compound over decades share a second feature, and it is the one that is hardest to imitate: they leave the operating businesses alone. Constellation is aggressively decentralised, and Leonard has been explicit that this autonomy is a large part of why owner-managers are willing to sell to them at all.

This is not a soft cultural preference, it is what makes the allocation model work. Central capital allocation only functions if the centre is small enough to stay objective and the businesses are self-sufficient enough not to need it. A head office that gets involved in operations becomes expensive, slow, and emotionally attached to the businesses it is supposed to be assessing dispassionately.

The failure mode is a centre that adds cost without adding return: a layer of reporting, some shared services nobody asked for, and a set of decisions made further from the customer than they used to be. That is the ordinary version of a conglomerate and it is why most of them destroy value. The version that works is a centre that does almost nothing except allocate capital extremely well and hold a very short list of non-negotiables.

The trap on the other side

Having argued for deploying deliberately, the opposite error is at least as common and considerably more expensive. It is deploying because there is money available.

A good year produces cash, cash produces pressure, and pressure produces a deal that would not have cleared the bar in a lean year. Nobody experiences this as lowering their standards. It is experienced as being ready, as the market being competitive, as this one being strategic. The word strategic in an investment discussion almost always means the numbers did not work.

The defence is procedural rather than intellectual. Write the hurdle down when you have nothing in front of you. Require the case to be made against the boring alternatives, specifically against paying down debt and against fixing something in the business you already own. And be willing to do nothing for a year, which is the hardest of the three and the one that most often separates the good long-run records from the average ones. The full argument for why not having to sell changes what you can buy is in permanent capital versus the fund clock, why consolidation usually fails anyway is in why roll-ups fail, and what has to be true before I will deploy against a target at all is on the buy page.

Hurdle rate tiers, the magnetic-hurdle observation and the decentralisation rationale are drawn from Mark Leonard's shareholder letters for Constellation Software and the company's published discussion of how it measures capital deployment.