How we scope a software project (and why fixed quotes fail)

Ask five agencies to quote the same project and you will get five numbers that differ by an order of magnitude. None of them are lying, exactly. They are pricing five different imaginary projects, because the real one has not been defined yet. A fixed quote on an unscoped project is a work of fiction — and someone always ends up paying for the rewrite.

Why estimates go wrong

Software estimates fail for a boring reason: the expensive parts of a system are invisible at the brief stage. The brief says users log in. It does not say what happens when someone needs access revoked mid-contract, or when two people edit the same record, or when the data import contains eleven years of inconsistent spreadsheets. Those cases are the actual project. The login form is a rounding error.

What we do instead

Every engagement starts with a paid discovery phase, one to three weeks depending on scale. We interview the people who will use the system daily — not just the people buying it. We map the workflow as it exists, including the workarounds nobody admits to in the kickoff meeting. Then we write it down: user journeys, a data model, integration points, and an explicit list of what version one will not do.

Discovery ends with three artefacts: a scoped specification, a phased delivery plan with real budget ranges per phase, and a prioritised risk register. Clients own all three outright. Some take them to another agency — genuinely fine with us. A well-scoped project is cheaper to build anywhere.

Ranges, then milestones

After discovery we quote in ranges that narrow as uncertainty resolves, and we bill against shipped milestones — working software you can use, not percentage-complete reports. If a milestone lands early, the next estimate tightens downward. If we discover a swamp, you hear about it that week, with options, while the decision is still cheap.

The uncomfortable question

If an agency offers a precise fixed price from a two-page brief, ask them which assumptions the number depends on. Either they cannot tell you, or the contract's change-request clause is where the real pricing lives. The cheapest quote and the cheapest project are rarely the same document.

Scoping is not bureaucracy before the real work. It is the part of the work where mistakes cost hundreds of pounds instead of tens of thousands. We would rather spend your first budget week there.