Most software companies pick a market, build a product, and try to sell it. GVP works the other way around.
We start with a person — usually someone running a business with $1M–$50M in revenue who’s been duct-taping their tools together for years — and we ask them what they wish existed. Then we go build it with them.
”Customer” is the wrong word
The word “customer” implies a transaction: you have money, we have software, we exchange. The relationship ends at the invoice.
A partner is different. A partner sits in the room while you’re sketching the data model. They tell you which automation will make their week and which one will silently break their workflow. They review screens before they ship. They’re on the hook with us — and we’re on the hook with them.
That’s how the work gets shaped right.
Three kinds of engagements
We talk about this on the how-we-work page, but it’s worth restating here:
- Quick wins — small builds where the problem is obvious and the fix doesn’t need debate.
- Shared builds — a problem deep enough that we co-invest with the partner who brought it.
- Venture products — opportunities we believe in enough to fund ourselves. Partners who help shape these benefit first.
Different engagements, same operating principle: the people who do the work get to shape the tool.
Why now
Five years ago, this model didn’t work at the scale we work at. Custom software was too expensive. Off-the-shelf was too generic. Modern infrastructure and AI cut the cost of building real software by an order of magnitude — and that means a 20-table restaurant can now afford the kind of bespoke tooling that used to be enterprise-only.
The window is open. We’d rather walk through it with partners than push product at a market.
— Jeff