004-scope-before-estimate.md
Before I give you an estimate
A useful estimate starts with the decisions we have not made yet.
I am wary of an estimate that arrives before anyone has discussed what counts as finished. A request can sound precise and still contain several different projects. “Replace the reporting dashboard” might include rebuilding the data underneath it, or it might mean changing the interface while keeping every existing calculation.
Both are reasonable requests. They require different work. Putting a number next to the sentence before resolving that difference gives everyone something concrete to repeat and very little reason to trust it.
Find the uncertain part
Before estimating, I want to separate work we understand from questions we still need to answer. Building a familiar screen may be straightforward. Establishing whether the source data supports the screen may take investigation. I would rather explain that uncertainty while there is still time to change the plan.
Consider a dashboard that needs to reconcile two systems with different definitions of a customer. The first useful piece of work might be comparing a sample of records with the people who use them. That can tell us whether we need a mapping, a change to the source, or a decision about which system is authoritative. Drawing more screens will not settle it.
I like a scope that names the decision this first piece of work will support. It gives discovery a stopping point. It also lets the client decide whether the answer is worth the cost before committing to the rest.
Make room for a smaller answer
Scoping should leave open the possibility that less development is needed. Perhaps an existing report answers the question once someone fixes its inputs. Perhaps a manual step is acceptable at the current volume. Those options belong in the conversation even when I would enjoy building the replacement.
That can be uncomfortable when the first conversation was about delivery dates. I still think it is better to spend a little time agreeing on the problem than to negotiate an increasingly precise price for different ideas of the solution.
An estimate becomes useful when the client can see what it includes, what remains uncertain, and what would cause us to revisit it. Then a change in scope can be discussed against the same written assumptions that produced the number.