Use completed projects to challenge a new cost estimate

A detailed supplier estimate can still miss costs that repeatedly arise during implementation. Alongside the current build-up, examine completed projects that resemble the proposed work. Compare what was estimated at a similar decision stage with the eventual cost and delivery date.

The UK Treasury's 2026 Green Book addresses optimism bias using historical forecast errors and comparable proposals. Those are public-sector appraisal instructions, not a mandatory budgeting rule for private companies. The useful business application is to challenge an internal estimate with evidence from outcomes already observed.

Preserve the original estimate in each historical case. Comparing final costs with a budget repeatedly enlarged during delivery can hide the error management wants to understand. If the first approval record cannot be recovered, mark that limitation rather than treating the latest budget as the original forecast.

Define the criteria before selecting examples. A small configuration change, a first installation and a full replacement may have different uncertainty even when all use the same software or equipment. Include unsuccessful projects where their records meet the criteria. Excluding difficult outcomes because they look unusual can bias the comparison.

Match estimate maturity. A concept estimate made before design should not be judged against the accuracy of a fixed quotation obtained after testing. Record whether scope, supplier selection and technical requirements were settled when each estimate was prepared. These differences can explain more than a broad industry label.

Homes England's reference-class guidance bases the approach on actual results from comparable projects and notes the importance of the base estimate. Use it as methodological context while choosing evidence suitable for this company's work. Do not import published sector uplift percentages into an unrelated business case.

Separate scope additions from cost growth on the original requirement, while retaining both in the overall outcome. A larger delivered project may justify additional spending, but repeated scope omissions can themselves be relevant evidence about the estimating process. Decide which question the comparison is intended to answer.

Normalize currencies and price dates with a documented method where those differences are material. Keep the original amounts available. An adjustment should make cases more comparable, not remove every unfavorable result. Have an independent reviewer challenge exclusions or changes that materially improve the apparent historical performance.

A handful of internal projects can expose missing cost categories without supporting a precise probability distribution. Describe the sample and its limits. Do not present a percentile from a thin or selectively documented history as a reliable chance that the next project will stay within budget.

Present the current estimate beside the historical comparison and identify why the new project should differ. Support a lower allowance with specific changes, such as completed technical testing or a narrower agreed scope. A general claim that the team has learned from the past needs an observable change behind it.

Check for overlap with contingency already included in supplier quotes or the risk analysis. Applying a historical uplift to a total that already covers the same risks can double count the allowance. Reconcile the base estimate, identified risk amounts and proposed overall budget.

Keep the comparison file with the approval and update it after completion. The next review then has another documented case with a known starting point, scope and outcome. That makes the evidence progressively more useful without turning one project's experience into a universal percentage.