Review the assumptions behind a work estimate

An estimate is difficult to compare when contributors are imagining different work. One person includes investigation and review; another counts only editing a known file. Examine the assumptions behind an effort range before using it to plan a deadline. The useful result is a conditional explanation, not a falsely precise promise.

Define a visible finished result

Break one deliverable into observable work items and describe what done includes. Specify the artifact, review requirement, and any acceptance condition. Identify work deliberately outside the estimate. Avoid using a broad label such as update the handbook when people could interpret it as anything from correcting a link to replacing the guidance.

Fictional practice example: a handbook update might involve three existing pages, or it might require investigating a new source before rewriting them. Use this as a planning exercise rather than a claim about a customer's delivery. State which scenario the first estimate addresses and leave the uncertain scenario visible.

Ask for ranges and their assumptions

Have contributors write plausible effort ranges independently before discussing them. Ask what information, access, and skill each estimate assumes. Record whether review and revision are included. Numbers without those conditions give the group little to examine beyond confidence or seniority.

In the example, one contributor may assume approved source material exists and only wording needs revision. Another may expect to locate an owner, reconcile contradictory guidance, and wait for review. Both estimates can be reasonable for their assumptions. The first task is to find out which work description matches the intended deliverable.

Explain disagreement before averaging

Compare the scope and assumptions responsible for the largest differences. Ask contributors to point to the work item that changes their range. Distinguish a different judgment about familiar effort from uncertainty about whether an entire activity is required.

Do not immediately average estimates based on incompatible definitions of done. Reconcile the work description, then ask whether each range changes. Preserve a real disagreement when evidence is missing. Choosing the middle number can look decisive while hiding the very condition that will determine whether the work fits the plan.

Separate active work from waiting

Describe the effort needed to perform the work separately from elapsed time dependent on another person's response. Identify the required input, its owner, and whether that owner has agreed to provide it. A short edit can still have a long delivery path when a necessary decision is unresolved.

Choose a small learning action for the unknown that most changes the range. In the fictional update, inspect the existing sources before estimating a full rewrite. Give that investigation a clear question and stopping point. It should reduce uncertainty, not quietly become the entire project under a different name.

Publish a conditional estimate note

Record the range, included work, assumptions, waiting dependencies, and the event that would require revisiting it. State who will obtain the missing information. If a delivery date is requested, explain which dependencies prevent translating effort into a reliable date yet.

Test the note by asking another person what would make the estimate no longer apply. If they cannot name a condition, the explanation is too vague. Revisit after the learning action and preserve why the estimate changed; a revision based on new scope is different from an unexplained missed commitment.