Founder's 90-day runway. The integration offer that would have cost him his roadmap.

The email arrived at 4:47 on a Friday, when the week was already over and the reasonable response was to let it sit until Monday. I opened it anyway because the subject line named a company I'd pitched eleven months earlier and written off as a dead lead.

The ask was enormous. A procurement manager wanted our core workflow engine embedded into their internal tooling across three continents. Enough money to cover six months of runway, maybe more, and the kind of logo that makes the next five conversations easier. The deadline was Monday. The catch was hiding in paragraph two: they needed us to rebuild the product as a white-label utility serving their specific data model, their specific approval chains, their specific reporting structure.

That email is the cheapest money a founder will ever be offered, because the invoice is paid in product direction rather than cash.

The most expensive question in the room

I knew the answer was yes before I finished reading. We had ninety days of runway. I'd spent that week doing the arithmetic founders do in private, working out which contractors to let go and which features to cut. This email was a door opening. The sensible thing was to walk through it.

What gave me pause was the word "embed." Not integrate. Embed. They wanted our product to become a feature inside their system, invisible to their end users, doing exactly one job exactly how they specified it. The version of us that would be real in a year was the version that had spent that year building a single client's plumbing.

I called a friend who had taken a similar deal two years earlier. He described the shape of his company in the present tense and I recognized it: the integration had become the product, and the product he'd originally built was a slide in a deck he no longer presented. He was profitable. He was also doing custom work for three clients that all wanted different versions of the same thing, and his roadmap was a set of requests from people who weren't his customers anymore.

The signal I kept circling: every integration request is a bet that the client's problem is your future market. Sometimes it is. But the scope of the request told me they were not buying what we'd built. They were buying our team's ability to build what they didn't want to build themselves.

The question that decides it

I sat with a blank document and forced myself to answer one question: if this client disappeared tomorrow, would the work we did for them help us build something we wanted to own?

The distinction isn't about customization. Every serious product does some amount of client-specific work, and the right integration can teach you things about your space you wouldn't learn any other way. The real test is whether the work accumulates or ends. Does the integration live in a part of the product you were already strengthening, or does it fork the architecture into a shape that only one customer will ever need? Does it compress your roadmap or replace it?

Runway math pulls hard in the other direction. Six months of runway is not abstract in a real company. It's the difference between hiring that engineer and extending the runway, between taking the overseas contract and protecting the roadmap, between shipping the AI demo and validating demand. I understand why so many founders take the money. The company you save today is the company you get to keep.

What I actually did

I said no to the white-label rebuild and counter-offered on the parts that fit where we were already going. That was the live part of the negotiation, and the client said yes to enough of it to matter. We did a scoped integration that shared the roadmap's bones, not a fork of it. The money was less, a third of what the original ask implied. It still bought us breathing room, and the work pointed where we'd already pointed ourselves.

The pattern I keep seeing across the founders I talk to, from Accra to Berlin to San Francisco, is the same one I nearly walked into: a large integration request reads as product validation and usually signals the opposite. A company with money is asking you to become their vendor. Vendors get paid well. They also get shaped.

The question I'd put to anyone holding an email like that on a Friday evening: what does your product become if this is the only client you ever needed to satisfy? If the answer is a company you wouldn't want to run, then the right move is rarely the obvious one. Ask what they actually need that overlaps with what you're already building, and price the rest accordingly. Sometimes the cheaper deal is the one that still leaves you owning your own roadmap.