The internal tool that wants to be a product.

It happens to every team that ships. You build a small internal thing to scratch your own itch — a dashboard, a script that became a service, an admin panel that quietly got good. One day someone looks at it and says the dangerous sentence: we should sell this.

Sometimes they’re right. Some of the best products in the industry started as a company’s internal plumbing that escaped. The pull is real, and I don’t want to talk you out of all of it. I want to talk you out of the version where you say yes for the wrong reason.

Here’s the trap. The internal tool is good because it was built for exactly one customer — you. Your constraints, your vocabulary, your weird edge cases, your willingness to tolerate a rough edge because you know where the bodies are buried. That fit is the whole reason it feels so sharp. And it’s exactly the thing that doesn’t transfer. A market is not you. The first external user doesn’t share your context, won’t forgive your rough edges, and needs the three explanations your team never wrote down because nobody on your team ever needed them.

The deeper cost is the one the growth story hides: shipping it externally means running a second team. Support tickets from people you’ve never met. Documentation for behavior you understood intuitively. Billing, uptime promises, a security review, a backlog now set by strangers’ needs instead of yours. The internal tool was a side effect of your real work; the product is real work, with its own gravity, pulling people off the thing you actually set out to build. You don’t get to keep the lightweight version. The moment it has outside users, it has outside obligations.

There’s also a quieter failure. The tool got good through a brutally tight feedback loop — the only user was in the room, so every rough edge got filed off within a day. That loop is what made it sharp, and it’s the first thing that breaks when the users aren’t you. Growth doesn’t preserve the thing that made the tool worth spinning out; it removes the conditions that produced it.

So before you let it escape, make a few things true. External pull you didn’t manufacture — people outside the building asking to pay, not you imagining that they would. A buyer who isn’t you — name a real person, in a real role, at a company that isn’t yours, with a budget. A willingness to staff it as its own thing, with its own owner, not as a nights-and-weekends side quest of the team that made it. And plain honesty about centrality: is this close to the core of what your business is for, or is it a fascinating distraction that happens to work? (The same core-versus-commodity question that governs build-vs-buy applies here, pointed the other way.)

The reversible move, if you’re tempted: don’t spin up a team. Expose the tool to one or two real external users as a thin, deliberately unofficial pilot, and watch what breaks. You’ll learn in a month whether there’s a product here — and you’ll have committed almost nothing. Keep the door swinging until the pull is strong enough to walk through on its own.

The test isn’t whether the tool is good. Of course it’s good — you built it for the world’s most demanding, best-informed customer. The test is whether anyone else lives in that world with you.