Operations & Delivery

Lesson 1 · Planning

Estimation & planning

An estimate is a forecast, not a promise — and a single number hides the uncertainty.

Three-point estimating asks for an optimistic (O), a most likely (M) and a pessimistic (P) value. The PERT estimate, (O + 4M + P) ÷ 6, weights the middle and admits the tail. Software estimates are skewed: tasks rarely finish much early and can run very late.

Projects are constrained by scope, time and cost — the iron triangle. Fix all three and quality is what gives, so decide in advance which one can flex. The critical path is the longest chain of dependent tasks. It sets the earliest finish date: a day's slip on it is a day's slip for the project, while tasks off it have slack.

Beware the planning fallacy: people underestimate their own projects even when they know similar ones overran. Counter it with reference-class forecasting — how long did the last five similar projects actually take? — and by giving ranges with a confidence level. Re-forecast as you learn. An updated estimate is not a failure.

Picture it

A team quotes eight weeks for a payments integration. The manager asks for three points: 6 weeks optimistic, 8 likely, 16 if the partner's sandbox is as unreliable as rumored. PERT gives 9 weeks. She checks the last three integrations: each took about 40% longer than first estimated. She tells stakeholders “nine to twelve weeks, and we will know more after the sandbox spike in week two”.

Key formulas & frameworks

PERT estimate
(O + 4M + P) ÷ 6
Iron triangle
Scope · Time · Cost — fix two, flex one
Critical path
The longest chain of dependent tasks
Reference class
How long did similar projects actually take?

Try it — Three-point estimate

PERT estimate

9.0 wk

Standard deviation

1.7 wk

Likely range (±1 SD)

7.3–10.7 wk

Above most likely by

1.0 wk

The band runs from optimistic to pessimistic, the thin line is the most likely value and the thick line is the PERT estimate. A long pessimistic tail pulls the estimate above “most likely” — which is why single-number estimates run late.

Practice

Score: 0 / 5

Question 1

Q1.A task is estimated at 4 days optimistic, 6 most likely and 14 pessimistic. What is the PERT estimate, in days?

Question 2

Q2.A task that is not on the critical path has five days of slack and slips by three days. What happens to the project's end date?

Question 3

Q3.Fixing scope, deadline and budget at the same time leaves no risk to quality.

Question 4

Q4.What is reference-class forecasting?

Question 5

Q5.Your team's past projects ran 40% over their first estimates. The first estimate for a new project is 10 weeks. What is the adjusted forecast, in weeks?

Up nextFlow, WIP & cycle time

Next lesson