The List: How We Decide Who to Contact

In the first post we made a claim that's easy to say and harder to live by: that for a small, senior shop, relevance beats volume — and that the work of relevance happens before you write a single word, in deciding who's even worth contacting. This post is about that decision. Not the secret sauce — we'll keep some of that — but the logic, because the logic is the part worth defending.

The short version: we don't build a list of who to contact. We build a list of who to eliminate, and whoever survives is the list. Almost all the effort goes into disqualifying.

Why disqualification, not discovery

Most outreach starts with a question like "who could be a customer?" — and the honest answer to that is "almost anyone," which is exactly why it leads nowhere useful. A list built on "could" is just volume wearing a nicer jacket.

We start from the opposite end. The default assumption is no, and a company has to earn its way onto the list by surviving a set of filters. This inverts the economics: instead of spending our scarce time writing to many marginal candidates, we spend it reading a few promising ones closely enough to be sure. The filtering is slow. What comes out the other side is small, and that smallness is the feature, not the bug.

What we actually look at

We look at signals an engineer can read, rather than the demographic guesswork that drives most "ideal customer profile" exercises. Three categories do most of the work.

Does our kind of problem plausibly exist here? We're good at a specific class of problem — the kind that lives in aging or strained backend systems, complex domain logic, platforms that have grown faster than their architecture. So the first question is whether there's any sign of that class of problem, as opposed to a company that's perfectly happy with what it has. There are public, observable signals that point one way or the other. We won't enumerate exactly which ones we weight most heavily — that's part of what makes the method ours — but the principle is simple: if we can't see a plausible version of the problem we solve, the company drops out here. Most do.

Is there evidence of real friction, not just our wishful reading of one? It's dangerously easy to convince yourself a company has a problem because you'd like them to. So we hold ourselves to evidence: something observable that suggests genuine strain, rather than a story we've constructed. The discipline is to distinguish "I can imagine they struggle with this" from "there's a visible reason to believe they do." Only the second kind counts. This is the filter that catches our own optimism, and it's the one we're strictest about.

Can we reach a person who actually decides? A perfect-fit company we can't reach is, operationally, not a fit at all. At a large enterprise the technical decision-maker may be unreachable by any honest means; at a smaller one, often not. So reachability is a real criterion, not an afterthought — there's no point qualifying a company on every other axis if the message lands in a void.

The test that ends it

All of this collapses into a single practical test, and it's the one we actually use to make the call:

Can we write a specific, true first sentence about this company's situation — by hand, with no merge tags — that they would recognize as accurate?

If yes, they're a candidate. If we find ourselves reaching for something generic, something that would be equally true of fifty other companies, that's the signal that we don't actually understand their situation yet — and we don't contact them. The sentence is the proof of work. It's hard to fake, it can't be templated, and writing it forces us to have done the homework the whole method depends on.

That single test is, in a sense, the entire method compressed into one line. Everything upstream — the filtering, the evidence, the reachability — exists to make that sentence possible to write honestly.

Why we're comfortable showing this

A reasonable objection: doesn't explaining your filtering method help competitors copy it?

Not really, and the reason is itself a point about the approach. The method isn't a trick you can lift; it's mostly work — the slow reading, the honest disqualification, the refusal to send the generic sentence. Anyone can know that's the approach. Far fewer will actually do it, because doing it is tedious and produces a list that looks alarmingly short. The moat isn't the secret; it's the willingness to keep the list small when every instinct says to make it bigger.

Which is also why this is the version of "how we grow" worth writing down. The next post in the series will get into what we're actually measuring as these contacts go out — and, once there's enough of it to mean anything, what the data says.

More to come.