Analogous Estimating

Schedule technique · See it on the map

Runnable here

Estimate a task's duration or cost by pointing at a genuinely similar past task and scaling from what that one actually took.

When to use it

Reach for this when you have a comparable completed project or task to point at and need a number fast — early in planning, or for a rough order-of-magnitude estimate before detail exists to support anything more precise. It's the estimate you can produce in an afternoon from history you already have.

When to avoid it

Don't use it when the past project you're comparing to isn't actually comparable — different team, different technology, different scale — because the estimate then inherits a false sense of grounding it hasn't earned. And don't use it as the final estimate for a task the decision genuinely needs precision on; it's the least accurate of the four estimating techniques here precisely because it skips decomposition.

Steps

1. Find one or more completed tasks or projects that are genuinely similar in scope, complexity and context — not just similar in name.
2. Record what the analogous work actually took, from real historical data rather than someone's memory of it.
3. Adjust for the specific differences between the analogous work and the task at hand (bigger, smaller, unfamiliar tech, different team) and state the adjustment, not just the final number.
4. Record the estimate along with which analogous task it was based on and what was adjusted, so it can be checked later.

What it produces

• A duration or cost estimate for the task, with the analogous reference and the adjustment reasoning recorded alongside it.

Common pitfalls

• The comparison task is chosen because its number is convenient (fits the target date) rather than because it's actually the closest match.
• The adjustment step gets skipped, so real differences between the two projects (a junior team versus the senior team that did the original work, say) never get reflected in the number.
• The estimate is presented with more confidence than the method supports — a single analogous data point dressed up to look like a calibrated estimate.

Worked example

A team scoping a new internal-tools migration recalls that last year's CRM migration, comparable in data volume and integration count, took eleven weeks. They estimate this migration at roughly ten weeks, adjusting down slightly because two of the four integrations this time are ones the team has already built connectors for.

Source

• PMBOK-6 §6.4.2.2

Where it comes from: this technique is named by the PMBOK Guide, 6th edition.