Five years ago this month, we signed the lease on a small office on Beaver Street and incorporated Semperr. A friend — an honest one, the rare kind — asked us over dinner what would keep the firm from turning into every other software company he knew. I gave him an answer at the time. I have thought about that answer at least once a week since.
The answer was that we would build things that worked the first time, for people who noticed. We would remain small enough to sign every deliverable. We would not raise outside capital. And we would hire only people who cared more about the work than the appearance of the work.
It was not a strategy. It was a set of refusals — four of them — and I was not at all sure they would hold. Five years in, I can report that they have, and I can also report that holding them has been harder and more interesting than any of us expected.
The cost of remaining small
A firm our size gets approached, roughly monthly, by one of three parties: a larger consultancy that would like to acquire us; a private-equity firm that would like to "roll us up"; or a venture investor who believes, with touching conviction, that we could be ten times bigger if we simply tried.
The polite answer to all three is no. The honest answer is that the size of the firm is not a limitation we are trying to overcome — it is a feature we are trying to protect. At forty-two people, we can do what a firm of four hundred cannot: sit in a room together, agree on what "correct" means, and ship to that standard without the work being diluted on the way out the door.
Growth is not free. Every additional person is a chance to lose a little of the standard; every additional office is a chance for a sub-culture to form that does not know why the standard exists. We would rather be the right size for the work than the right size for the pitch deck.
What has not changed
The things on the original list — the four refusals — are all still on the list. Some of them are harder than we thought. Answering the phone, for instance, is a commitment that scales with the firm, not with the product, and it has cost us more hours than any two engineering problems combined. But it is also the single thing our clients mention most often when asked why they stayed.
Not raising outside capital has been the easiest refusal to keep and the most consequential. It means we can say no to work that would pay well but compromise the standard. It means a bad quarter is our problem, not someone else's board meeting. It means the people who do the work own the firm, and when we make money, they make it.
A note on compounding
What I did not understand at the beginning, and understand now, is that small, honest technical decisions compound. A Trace report that has always been right is easier to trust on the thousandth Friday than it was on the first. A Clad citation that has never been wrong gets cited by memory. A Dev engineer who has always picked up the phone is the engineer the client asks for by name.
This is not romance. It is a plain fact about the economics of trust. Firms that spend their first five years accumulating trust do not need to spend the second five purchasing it.
The same answer, then
So when people ask now what keeps us from turning into every other software firm, I give them the same answer I gave my friend five years ago. The answer was not clever. It was not a strategy. It was four things we decided we would refuse to do, and it has turned out, against some odds, to be enough.
We will be here in another five. We intend to still be forty-something people, still answering the phone, still picking up after ourselves, still signing our names to the work. If you would like the next dispatch in your inbox, the list is at the bottom of the page.
— Eleanor