Skip to main content
Back to module
Project Management Fundamentals · Planning

Estimating and budgets

Reading39mLesson 3 of 3 in this module

Make three-point estimates of effort, turn them into a simple budget and add a sensible contingency.

Why it matters

Every sponsor asks two questions: "When will it be done?" and "How much will it cost?" Answering with one confident number ("12 days") hides the uncertainty that every estimate has. When reality turns out different, the team looks unreliable. Good estimating is not about predicting perfectly; it is about being honest about the range and planning for it.

Effort vs duration

  • Effort is the amount of work, in person-hours or person-days: "the form takes 24 hours of work".
  • Duration is calendar time: if Priya can spend 4 hours a day on it, 24 hours of effort takes 6 working days.

Mixing these up is a classic mistake. People are rarely 100% available; meetings, other projects and support work take time. Many planners assume 5 to 6 productive hours in an 8-hour day.

Ways to estimate

  • Expert judgement: ask the person who will do the work. Usually the best starting point.
  • Analogous: "the last site like this took 3 weeks." Quick, but only as good as the comparison.
  • Parametric: use a rate. "Each page takes about 6 hours to build; 4 pages = 24 hours."
  • Bottom-up: estimate each WBS work package and add them up. Slower, but the most accurate and easiest to defend.

Three-point estimates

Instead of one number, ask for three:

  • O (optimistic): if everything goes well.
  • M (most likely): the realistic case.
  • P (pessimistic): if things go wrong, but not disastrously.

Then combine them. Two common formulas:

Triangular average:   E = (O + M + P) / 3
PERT (beta) average:  E = (O + 4M + P) / 6
PERT spread:          SD = (P - O) / 6

PERT weights the most likely value more heavily. A worked example for the sign-up form:

O = 16 h, M = 24 h, P = 44 h
Triangular:  (16 + 24 + 44) / 3     = 28 h
PERT:        (16 + 4 x 24 + 44) / 6 = 156 / 6 = 26 h
Spread:      (44 - 16) / 6          = about 4.7 h

Notice the expected value (26 h) is higher than the most likely value (24 h). That is normal: tasks can overrun by much more than they can come in early, so the average drifts later. A single "most likely" number is too optimistic on average.

From effort to budget

A simple project budget has:

  1. Labour: effort per work package times a rate. Even volunteer or internal time has a cost worth tracking.
  2. Non-labour costs: software, hosting, equipment, materials, travel, training rooms, permits.
  3. Contingency reserve: money (and time) for known risks, the "known unknowns" in the risk register. Often calculated from the risks, or set as a percentage, commonly somewhere around 5 to 15% on small projects depending on uncertainty.
  4. Management reserve: held by the sponsor for "unknown unknowns". It is usually outside the project manager's control.

A worked example for the food bank site:

Labour   PERT total 150 h x USD 20/h (contractor rate)  = 3,000
Hosting  existing host, domain renewal                  =    20
Testing  2 second-hand phones for testing               =   280
Training room and printing                              =    50
Subtotal                                                = 3,350
Contingency (about 15%, design and photo risks)         =   500
Total project budget                                    = 3,850

That fits under the USD 4,000 constraint from the charter, with a small margin. If it had not, the options are the usual trade-offs: cut scope, find cheaper resources, or ask for more money.

Tracking costs

During the project, compare planned cost to date with actual cost to date and with how much work is actually done. Spending exactly half the budget means little if only a quarter of the work is finished. More formal techniques such as earned value management build on this idea.

Common mistakes

  • Estimating for someone else without asking them.
  • Forgetting overhead: meetings, reviews, testing, fixes, handover.
  • Padding every task secretly instead of a visible contingency.
  • Giving a single number when the range is wide. Say "about 4 weeks, likely between 3.5 and 5."
  • Never re-estimating. Update estimates as you learn; the early ones are the least accurate.

Quick recap

  • Effort is work hours; duration is calendar time. Account for availability.
  • Three-point estimates: optimistic, most likely, pessimistic; PERT uses (O + 4M + P) / 6.
  • A budget is labour plus non-labour costs, plus contingency for known risks and a management reserve.

Try it

Open the Do tab and produce three-point estimates and a one-page budget for your WBS.

Also in this lesson: