effort
Also called: person-days, person-hours
How much work a task takes, counted in people's working hours or days, as opposed to how long it takes on the calendar (duration). Two people do not halve the calendar time when the work cannot be split or needs time to hand over.
Effort is the amount of work a task needs, measured in person-hours or person-days. Duration is the calendar time from start to finish, and it includes waiting, briefing and the parts only one person can do. The two answer different questions: effort drives cost and capacity (how many people-days do we have, what will they cost?), duration drives the schedule (when will it be done?).
Duration is never simply effort divided by the number of people. Most tasks have a part that can be split, a part that cannot (a final pass one person must hold in their head) and a coordination cost that grows with every person added (briefing, agreeing what "done" means, merging work). Some tasks are mostly waiting: replies, approvals, someone else's timetable. Waiting doesn't split; a second person just waits too.
So adding people shortens a task only up to a point and always raises its effort: you buy calendar time with extra person-days. That trade is worth it only on the critical path. Before adding anyone, ask four questions: Is the task critical? Can its work really be split? Who briefs the newcomer, and is that person critical too? Is the duration work or waiting?
When the answers say no, change the plan rather than the headcount: start something earlier, run work in parallel, or cut scope on the critical path.
Example
Vira's review of five curriculum units is on the critical path of Klasna's physics launch and takes 10 days. Leo suggests five reviewers and two days. The review breaks down into:
- 8 days that can be split (reading and checking the units);
- 2 days that cannot (Vira's final pass across all five units);
- 1 day per extra reviewer to brief them and agree what counts as an error.
Duration with n reviewers = 8 ÷ n + 2 + (n − 1) × 1:
Five reviewers are slower than two. With two, effort rises to 12 person-days (8 + 2, plus the briefing day for both Vira and the newcomer) for 7 calendar days: three days off the critical path, bought with two extra person-days and an assumed 8,000 UAH.
Waiting is the other case. Roman's tutor screening shows 9 days of effort but 18 days of duration, because candidates take days to reply and are free only in the evenings. A second recruiter would mostly wait twice; asking for free slots in the first call shortens it more.
Common mistakes
- Dividing effort by people. Duration includes parts that cannot be split, coordination and waiting.
- Adding people off the critical path. Days saved on a task with float move nothing and still cost effort.
- Forgetting who briefs the newcomer. If that person is on the critical path, the help can slow the project down.
- Treating waiting as work. More people don't shorten replies or approvals; start earlier or ask differently.
- Planning capacity in calendar days. Capacity and cost are counted in person-days of effort, the schedule in duration.
Learn it in the course
- Case: the deliverable list that doesn't fit · Twenty-five deliverables with effort ranges, 150 person-days of capacity. Add up effort by priority, challenge two false musts, and write Leo the scope decision with a drop order.
- Effort is not duration: when adding people helps · Why ten days of review don't become two with five reviewers: the parts that can't be split, the cost of coordinating, and the waiting no extra person removes.
- Break the work down from deliverables · Deliverables first, tasks second: a tree that covers all the work of the physics launch and nothing twice, flattened into a list with owners and effort ranges.
- Re-plan from today · Reforecast the remaining work from where the project stands, keep the range, let sunk effort stay out of the choice, and tell the sponsor what changed and what you need.
- Must, should, could: cut scope, not the date · When the list is longer than the time, sort every deliverable with one test, agree in advance what is dropped first, and keep the fixed date.