Capacity
Estimate task effort without pretending you know the exact answer
An effort estimate is a planning input, not a promise that reality will cooperate. Useful estimates are specific enough to reserve time and honest enough to show what could change them. This guide explains how to separate active work from waiting, use a small range privately, choose a working allocation, and learn from a completed task without building an elaborate estimation bureaucracy.
By QuadrantWorksUpdated 6 min read
The short version
- Estimate the work needed to produce an accepted result, not a vague topic or an entire elapsed week.
- Separate active effort, counterpart waiting, and transition time before choosing an allocation.
- Use assumptions and a discovery step when uncertainty dominates the estimate.
- Compare estimates with observed active work; do not treat elapsed task age as a time-tracking record.
In this article
Define what finished means before estimating it
Prepare the partner update could mean collecting three figures, drafting a paragraph, obtaining approval, or redesigning a presentation. Those are different effort commitments. Start with the accepted result: publish a one-page partner update using the agreed metrics, with the account lead's approval. Then list the small checks required to produce it. An estimate for a named result is easier to review than an estimate for a broad subject.
In this fictional example, Leena owns the update. She expects to gather figures, draft the note, review it once with the account lead, and publish it. A design refresh is explicitly out of scope. If the requester later adds that refresh, the estimate should change. Re-estimating after scope changes is honest maintenance, not evidence that the original number was careless. Keep the assumption visible where the commitment is discussed.
Separate effort from elapsed time and waiting
Leena may spend 90 minutes actively producing the update while the approval spans two working days. The two-day interval matters to the deadline, but it is not two days of her effort. Conversely, waiting does not make the next action cost-free: asking a precise question, reviewing the response, and incorporating a change can all require active time. Represent those activities without reserving an entire workday for someone else's reply.
Transition time is a third quantity. A 15-minute gap between sessions protects against adjacent commitments touching, but it is not necessarily additional task effort. If you count buffers in a weekly capacity worksheet, do not also add the same gap to every task allocation without a reason. Choose one consistent budgeting convention and explain it to the people reviewing capacity. This guide's examples keep effort and calendar transitions separate.
| Quantity | Fictional example | Where it belongs |
|---|---|---|
| Active effort | 30 min gathering figures and 45 min drafting | Task allocation or a private effort worksheet |
| Review effort | 15 min checking the account lead's response | The owner's active effort estimate |
| Elapsed waiting | Approval expected the following day | Dependency and checkpoint notes |
| Transition | 15 min before the next meeting | Calendar buffer or capacity allowance |
Use a range before selecting the working allocation
A private estimate can begin as a range: 75 to 120 minutes if the source figures are ready, longer if they need reconciliation. Explain the reason for the range rather than attaching unsupported confidence percentages. The low end assumes a clean input and one review. The high end allows a modest revision. A range without assumptions is simply two uncertain numbers instead of one.
Critical Path stores a single time allocation in 15-minute steps, not a statistical probability distribution. Leena can choose 105 minutes as a working allocation and write the assumptions in the execution note. That choice is a planning judgment, not a mathematically guaranteed percentile. If failure to finish would carry a serious consequence, use a more conservative allocation or create an earlier decision checkpoint rather than relying on optimistic interpretation.
Use a discovery step when the unknown dominates
Some work cannot be usefully estimated until a key fact is known. If Leena does not know whether the three metrics exist, multiplying a guess by two is not a reliable solution. Create a short discovery action: inspect the available metrics and identify any missing input. Define its output as a decision about the remaining work, not a promise to finish the entire update inside the discovery window.
Suppose a 30-minute inspection reveals that one metric needs another team's reconciliation. Leena can now separate a request for that input from drafting the sections already supported by evidence. The revised plan may still miss the original deadline, and that should be communicated. Discovery is useful because it exposes the constraint early; it is not a technique for making uncertainty disappear from the calendar.
- Name the missing fact that would materially change the estimate.
- Choose the smallest inspection or conversation that can reveal it.
- Define the discovery result: proceed, reduce scope, request input, or stop.
- Estimate the remaining work only after recording what you learned.
Check session shape, not just the total minutes
A three-hour task is not automatically feasible because three isolated hours remain somewhere in the week. Some activities need an uninterrupted interval; others can be separated by a review or a handoff. Write down the natural stages, such as inspect inputs, draft, and verify. Then ask which can be completed independently and which must remain together to avoid wasteful restarts.
The current Critical Path calendar model caps a task's session at two hours and places at most one block for that task per working day. A three-hour allocation therefore needs more than one working day under that model. Keep separate tasks for genuinely different deliverables or owners, not simply to bypass the scheduling boundary. If the result requires an uninterrupted three-hour session, recognize the mismatch and agree a workable approach instead of assuming the product will infer it.
Review observed effort without turning estimates into a performance weapon
After the fictional update is accepted, Leena records an approximate active-work total of 120 minutes: 30 collecting figures, 60 drafting, and 30 revising. Compare that with the 105-minute allocation and ask why the difference occurred. The extra revision might indicate unclear acceptance criteria, a missing source, or ordinary variation. A single task does not establish a stable productivity rate or justify ranking employees by estimate accuracy.
Do not substitute the time between task creation and completion for active effort. Critical Path does not claim to have measured every minute you spent on this task; elapsed lifecycle metrics include waiting and other activity. Use a short manual note or an approved time-recording process if observed effort is important. Review a few comparable completed tasks over time, and adjust your assumptions when a pattern is actually visible.
Keep a lightweight rule for the next commitment
A practical personal rule might be: define acceptance, identify the largest unknown, choose a range, then allocate a rounded working value. Spend more estimation effort only when the decision warrants it. A familiar internal message may need a quick judgment. A first-time integration review with several decision makers may need discovery and explicit assumptions before anybody should commit to a date.
At the weekly review, look for three recurring causes of variance: hidden scope, missing inputs, and underestimated review work. Improve the corresponding preparation rather than adding a blanket cushion to everything. Keep unused contingency visible instead of automatically filling it. The aim is a believable execution budget that supports better decisions, not a table of precise-looking numbers that nobody trusts.
Common questions
Should I estimate every task to the nearest minute?
No. Use a level of detail that helps the planning decision. Critical Path allocation moves in 15-minute steps; the meaningful work is clarifying scope, assumptions, and whether an appropriate session exists.
Does an allocation guarantee the work is scheduled?
No. An allocation describes effort. Calendar placement also depends on dates, working windows, existing events, buffers, and the scheduling model. Review actual placement and unresolved constraints.
What should I do when the estimate is mostly unknown?
Create a bounded discovery action with an explicit learning result. Use that evidence to estimate or renegotiate the remaining commitment, and communicate any deadline risk early.