The answer is that the same contract that buys six months of runway can quietly turn a product company into an agency, and the line is thinner than the founder expects. A services contract pays for the roadmap, but it can also become the roadmap, one renewal at a time.
I got the email on a Friday evening, 18:42, which I remember because I was still in the office and the cleaning staff had already started on the floor below. It was from a large company, the kind of client I had been chasing for a year. They wanted a custom AI workflow tool. The budget number was transformative: six months of runway.
I drafted the acceptance in my head before I hit the end of the second paragraph. Then I stopped, because I had seen this exact moment go wrong for two other founders, and I had almost walked into it myself once before.
The contract that pays for the roadmap
The logic is seductive. You have a product with real users but slow revenue. You have a burn rate that makes sleep difficult. Then an email arrives offering you the entire runway problem solved, in exchange for a few months of your team's time building something the client wants.
I went through the offer line by line. The build was scopeable. The timeline was realistic. The payment schedule was a third on signing, a third on delivery, a third on approval. I had said yes to worse terms before.
The problem was never the terms. It was what happened to the team while the work was in flight. Your two engineers cannot build the client's custom workflow and your product roadmap at the same time. They can try, and for two weeks it feels like they are managing. Then a demo for the client slips, and suddenly the client's deadline is the priority, because the client is paying and the product's users are not.
I had watched a founder named Kofi go through this. His payments API had two enterprise pilots waiting on a feature he had promised for a quarter. He took the custom integration contract to cover payroll for the quarter. The feature shipped five weeks late. One pilot left, and the other put their deployment on hold. He had bought himself runway and spent his product's momentum to do it.
Where the line moves
The mechanism that turns a product company into an agency is not the first contract. It is the second one.
The first custom build creates a referenceable client, a proof point, and a growing invoice. When that client comes back with a follow-up request, or when a similar company approaches you because they heard about the first project, saying no gets harder. The revenue is known. The work is familiar. Your engineers already know the codebase.
I could feel the pull in my own inbox that Friday. The client had already mentioned "a few related things we'd love your input on." That sentence was the trap, and it was a kind sentence, the most dangerous kind. I knew that if I said yes to the build, the related things would become the second contract, and the second contract would set the pattern where our roadmap became whatever our clients wanted us to build next.
[Here is where the scene turns. The named founder returns, and the decision resolves just in time.]
I called Kofi before I replied to the email. He asked me one question I still think about: "When was the last time you shipped something a paying user asked for, rather than a client?"
I could not answer it fast enough.
So I did something that felt wrong for a founder with a burn rate. I asked the client for a narrower scope, a smaller project that did not require my engineers, and I quoted a delivery timeline that protected a specific block of product work we had already committed to. I told the client what I was protecting and why. I did not say yes to six months of runway. I said yes to a month of it, with my team's product time walled off.
The client said fine. That was the part that kept me up. I had been so ready to bend over backward for a contract that the client themselves treated as replaceable.
The capability you are actually selling
What I understood that weekend is that a founder's most expensive asset is the team's focus, and every services contract is a bet against it. The contract is not the enemy. The unexamined contract is the enemy.
The question to ask is not "can this pay the bills?" It is "what does saying yes permanently change about what this company builds?" If the answer is the roadmap, you have not bought runway. You have sold the company's direction, and you have priced it at whatever the monthly retainer happened to be.
A product company can take services work and survive. It happens all the time. But it survives as a product company only when the founder draws the line in advance, names the product work that is off limits, and is willing to walk away from a client who will not respect it.
The Monday after that Friday, I moved the product roadmap to the top of the office whiteboard and wrote the protected block in permanent marker. The client's project went on a separate line underneath. It was a small gesture. But every time an engineer asked what was on the board that week, the answer told them which company we were.
Comments
No comments yet.