Specs
The “serve everyone” product breaks the moment real data shows up
Serving everyone sounds like a bigger market. What it usually produces is a product that's secretly bespoke for every customer, breaks the first time real data arrives, and can't be built twice the same way.
A pattern shows up at a lot of fast-growing startups, and someone described it sharply from the inside: the product is supposed to serve everyone, it works great in the demo, and then it shatters the moment a real customer's actual data hits it. So the team patches it for that customer. Then the next customer's data breaks it differently, and they patch that too. Six months in, you don't have one product — you have a dozen bespoke ones wearing a trench coat, and every new logo makes it worse.
The instinct behind "serve everyone" is reasonable: a bigger addressable market, more deals you can say yes to, no painful narrowing. But "serve everyone" isn't a scope. It's the absence of one. And the absence of scope doesn't expand your market — it just moves the cost from the sales conversation to the engineering team, where it compounds silently.
Why real data is the moment of truth
A demo runs on data you chose — clean, shaped to fit, edge cases quietly absent. Real customer data is the opposite: messy, contradictory, full of the cases you never specified because you never decided they existed. The product doesn't break because the code is bad. It breaks because nobody ever decided what the product is supposed to do when the data looks like that — and "serve everyone" guarantees the data will eventually look like everything.
Every unhandled case becomes a per-customer patch. Every patch makes the next change riskier. Eventually nobody can describe what the product does without naming a specific customer, which is the clinical definition of having no product at all.
The fix is a scope you're willing to defend
- Pick the customer you're actually for. Not the biggest imaginable market — the specific buyer whose data and workflow you'll handle deliberately, completely, and better than anyone. Depth for one beats breadth for none.
- Decide the edges on purpose. The cases real data will throw at you — the empty field, the duplicate, the wrong format, the volume spike — aren't bugs to patch later. They're product decisions to make now. A gap analysis on the spec is cheaper than a war room in production.
- Make "no" a feature, not a failure. Every customer you deliberately don't serve is a customer who can't drag your product into a one-off. Saying no to a use case is how you keep the product buildable twice.
- Write down what's in and what's out. A scope that lives only in the founder's head can't stop a sales rep from promising the moon. A scope on paper can.
What the "stay flexible" crowd gets right and wrong
Right: early on you genuinely don't know exactly who you're for, and over-narrowing on day one can cut off the customer you'd have been great for. Some flexibility is how you find the niche.
Wrong: flexibility and "serve everyone" aren't the same thing. Flexibility is staying open about who while being ruthless about what for whoever you pick. "Serve everyone" skips the picking entirely and calls the resulting chaos a market strategy. The teams that escape the bespoke-treadmill don't do it by working harder. They do it by deciding, on purpose and in writing, what the product is and isn't — before the next customer's data decides for them.
- “Serve everyone” isn't a scope, it's the absence of one — and the cost moves to engineering, where it compounds.
- Demos run on data you chose; real data exposes every case you never decided to handle.
- Unhandled cases become per-customer patches until nobody can describe the product without naming a customer.
- Pick a specific customer, decide the edges on purpose, and put what's in and out in writing.
Decide the scope before the data decides for you
Cadenly's Spec workflow turns “serve everyone” into a defined scope with the edge cases made explicit — and a gap analysis that catches the holes before real customer data does.
Start free →