How we decide which businesses to acquire.

Here’s the question we ask about every business we look at: would it be worth more with us than without us, once you count everything it takes to make it better? If we can’t answer that clearly, it’s probably not a fit, even if the business is great.

Two things have to be true. The business has to be good today. And we need a specific idea of how to make it better for the people who use it. One without the other usually isn’t enough.

In practice we look at four things at once: the business itself, how solid the numbers are, the price, and whether we have the people to run it well. Doing great on one doesn’t make up for a problem with another.

Right now we’re looking for subscription software making $5k–$20k a month in recurring revenue, that’s been around for three to eight years. Think developer tools, creator software, productivity apps and the simple tools small businesses run on.

The businesses we like most do one job well. Customers come back without being chased. They renew without a discount or a sales call, and they can tell you why they’d miss it. New customers keep showing up even when the founder stops posting about it.

We also like a business that already works but has one clear thing that could be better. We see three versions of this a lot. A good product where too many new users give up during setup. Loyal customers, but the business costs more to run than it should. Or people clearly value the product, but the pricing plans don’t match how they actually use it.

We’d rather see billing exports and bank statements than a dashboard. That’s not because we doubt anyone. Dashboards just leave things out.

For recurring revenue, we look at each customer contract. An annual plan paid up front counts as a twelfth of its price each month, not as one big month.

For retention, we follow the same customers over time. New sign-ups can hide the people who quietly left, so we want to see who stayed, who left and who moved up or down a plan.

Then there’s the part that surprises people most. Small businesses usually add the founder’s own pay back into profit. That makes sense on paper. But when the founder steps back, the work doesn’t go away. Someone still has to answer support, fix bugs and keep the servers up. So we count what it would cost to pay someone for that work.

That can change the picture a lot. Here’s a made-up example. A business shows $108,000 a year in profit. Pay someone to do the founder’s job and cover maintenance, and it’s closer to $54,000. After tax and investment, it’s about $40,000 in actual cash. Same business, same asking price, very different numbers. It’s better for everyone to see that early.

We also look at how much depends on a single customer, channel or platform. And at how many hours the business really takes each week, because somebody has to put those hours in.

A number we can check against the original records counts for more than one that only lives in a spreadsheet. And an unknown is just an unknown. It isn’t good news.

Most small software companies don’t have audit-ready books, and that’s completely normal. A gap in the records is something to work through together.

We keep two questions apart. The first is how good the business is today. Does the revenue check out? Do customers stick around? What does it earn after real costs? Is the code in decent shape, and can ownership actually transfer?

The second is what we’d actually change. A good answer is short: one to three specific things, each with a clear problem, a starting point, someone responsible, a cost and a way to tell if it worked. “We could raise prices” or “we could try ads” doesn’t count for much on its own.

Keeping these apart stops us from fooling ourselves. A strong business with not much to fix can still be a great long-term home. A shaky business with an exciting improvement story is usually a pass. We don’t want a deal that only works if we rescue it.

For us, price starts with the business as it is today, not with what we hope to turn it into. We do look at three versions: what happens if things go badly, the business as it runs now, and the business after our changes. The middle one carries the most weight.

The improvements are our work and our risk. If they don’t work out, that’s on us, so they shouldn’t drive what a business is worth as it stands.

There are also two limits. One is what the business is worth, based on the cash it can produce. The other is what we can afford while still looking after the products and customers we already have. A price has to make sense on both. Sometimes a great business just isn’t the right buy at its asking price, and that’s okay.

We also think through what could go wrong. Your biggest customer leaves. Sales slow down. More people cancel. A vendor doubles its price. Handing over the founder’s work costs more than expected, or our improvements take longer. In those cases, we care more about the worst month than the average one. And we look at the deal as if we’ll never sell it, because selling isn’t the plan.

Every business needs someone looking after it: someone to run it day to day, cover support and make the first improvements. A deal can work on paper and still fall apart because nobody has the time to do it well.

So we’d rather go slowly than take on more than we can handle.

The first job is to keep things steady. That means getting to know the numbers, reading through the support queue and talking to customers before changing much. After that come a few improvements, each with a clear reason and a way to check whether it worked.

And we look back, including at the things that didn’t work. That’s how we get better at this.