How We Scope a Project Before We Send a Proposal

The call comes first

Most agencies ask for a spec doc and send back a number. We do it backwards: a 45-minute call before anything gets written down. Half of that time is us asking questions the client hasn't thought about yet — who's the second user of this thing, not just the first; what happens if it succeeds and 10x's traffic in month two; what's already been tried and quietly shelved.

That call kills more bad scope than any amount of documentation would. If a client can't answer "what does done look like for version one," that's the actual first task, not the build.

What goes in the estimate

We break the work into weeks, not hours. Hour-based estimates give a false sense of precision on things nobody's built before. A week either has a shippable slice at the end of it or it doesn't — that's a number we can stand behind.

We also write down the assumptions the estimate depends on: which API we expect to integrate with, whether design is ours to do or already exists, whether there's a real staging environment or we're building one. When an assumption turns out wrong mid-project, that's a scope conversation on day three, not a surprise on day thirty.

The part clients are surprised by

We'll tell you if a project doesn't need us yet. If the honest answer is "validate this with a landing page and five customer calls before writing code," we say that, even though it means a smaller invoice or no invoice at all. It costs us short-term revenue and buys the only thing that actually gets us referrals.