Post-Investment Review Template (Capital Project Look-back)

Post-Investment Review: [Project name]

Post-investment review (PIR, also called post-completion audit or capital project look-back) comparing what a capital project actually cost, when it was delivered and what it achieved with what was approved: original approval baseline, delivery against cost, schedule and scope, benefits realised, financial re-appraisal with an NPV bridge, root causes, accuracy of the original assumptions, lessons, recommendations and close-out. Written as original practice guidance; public bodies in the UK evaluate under HM Treasury guidance (the Magenta Book for evaluation, with the Green Book 2026 edition for appraisal), which is cited by reference only. Replace every [bracketed] value and every row marked Example. This is a document format, not financial, investment, tax, legal or accounting advice.

Document control

Approvals

Revision history

Contents

1. Summary and overall rating
2. Scope and method of the review
3. Original approval baseline
4. Delivery against cost, schedule and scope
5. Benefits realised
6. Financial re-appraisal
7. Variance causes
8. Accuracy of the original assumptions
9. Lessons learned
10. Recommendations and actions
11. Close-out
12. Glossary and references

1. Summary and overall rating

Conclusion (Example): the project is delivering its run-rate savings and still clears the hurdle, but about two-thirds of the approved NPV was lost (297.6 of 437.2, 68%), mostly to cost overrun and a one-quarter delay (265.4), with a smaller run-rate shortfall from FY2030 (38.2). The decision to invest was sound on the evidence available in March 2027; the shortfall came from an under-scoped building survey and an installation plan that did not allow for the peak-season freeze. The main lesson for future appraisals is to schedule installation around operational freeze periods and survey the site before ordering.

2. Scope and method of the review

2.1 Questions answered

1. Was the project delivered to the approved cost, timing and scope?
2. Are the approved benefits being realised, measured against the recorded baseline?
3. What is the project's value now, on the original basis and looking forward?
4. Why did outcomes differ from the plan?
5. How accurate were the appraisal assumptions, and what should change in the appraisal and delivery process?

2.2 Evidence

2.3 Limitations

Example: FY2029 savings include three months of forecast (October to December). Savings are measured as the change in actual cost against the FY2026 baseline adjusted for volume (orders shipped) and wage inflation; volume grew 6% between FY2026 and FY2029, so an unadjusted comparison would understate savings.

2.4 Independence, evidence standards and tone

The value of a post-investment review depends on whether the approving body trusts it. Choose a reviewer who did not prepare the original appraisal or run the project: internal audit, FP&A or the finance lead of another business unit. The project team supplies data and comments on the draft, but the conclusions are the reviewer's. Record the reviewer's independence in Document control, and say plainly if a fully independent reviewer was not available.

Hold the review to the same evidence standard as an audit working paper. Agree costs to the general ledger and the assets under construction account rather than to the project tracker; agree savings to cost-centre actuals, adjusted for volume and price by a method fixed at approval; and support dates with signed acceptance certificates. Where evidence is incomplete, say so in section 2.3 and limit the conclusion accordingly instead of filling the gap with estimates from the sponsor.

Keep the tone factual. The purpose is better decisions, so separate the quality of the decision from the quality of the outcome: a sound decision can still produce a poor result because of an event nobody could reasonably have foreseen, and a careless decision can turn out well by luck. Name processes and assumptions, not people, and give the sponsor the chance to respond in writing before the report goes to the approving body.

2.5 Process

The review follows the same steps for every project, so results can be compared across the capital programme.

3. Original approval baseline

4. Delivery against cost, schedule and scope

4.1 Cost

Example checks: actual capex 1,650 + 520 + 365 + 300 + 125 + 50 = 3,010. Overruns before contingency: 100 + 55 + 120 + 35 + 50 = 360; contingency 150 absorbed part, net overrun 210. Tolerance was the lower of 10% (280) and EUR 250k, so 250; the overrun of 210 stayed within it.

Cash timing: EUR 2,610k paid in FY2027 and EUR 400k in FY2028, consistent with the deferral reported in Rolling Forecast (finance/rolling-forecast). The supplier invoiced 314 of the 476 acceptance milestone (MS-3) in December 2027 against the completed factory acceptance test; the remaining 162 was paid in Q1 2028 after site acceptance on 2028-03-10, and the retention (MS-4, 238) in Q2 2028 after final acceptance on 2028-05-15 (CON-001). Check: 162 + 238 = 400; 3,010 - 400 = 2,610.

4.2 Schedule

4.3 Scope and quality

Record any scope added or removed and whether the asset meets its specification. Example: contract throughput (9,000 order lines per hour) reached in the performance test on 2028-05-10. Safety guarding (WP-09) was added after the occupational safety inspection. No scope removed. Conditions CON-001 and CON-002 were met: the retention was paid after stable operation, and the two contingency draws above EUR 50k (WP-04, WP-02) carry CFO approval.

4.4 Interpreting delivery variances

Read the cost table at work-package level before looking at the total. A total within tolerance can hide a large overrun in one package offset by an underspend or by contingency; in the example the gross overruns before contingency were 360, almost two and a half times the contingency, even though the net overrun of 210 stayed within the 250 tolerance. Report both figures: the net overrun shows whether the approval limit held, the gross overrun shows how good the estimate was.

Separate three kinds of difference. Scope changes (work added or removed, such as the safety guarding in WP-09) are decisions, and the question is whether they followed the change process. Estimating errors (the building works) show whether the original estimate was robust. Delivery problems (the paused installation) show whether the plan and its controls worked. Each kind leads to a different lesson, so do not net them off against each other.

Treat timing in the same way as cost. A one-quarter delay rarely shows in the capital cost table, yet in the example it cost almost as much NPV (122.0) as the whole capital overrun net of phasing (133.4). Show the approved and actual dates of every milestone and the reason for each slip, and carry the effect into the benefits and the re-appraisal.

5. Benefits realised

Savings method (Example): baseline cost per order for FY2026 is inflated by actual wage growth and multiplied by actual FY2029 volume; the difference from actual cost, less the maintenance contract and energy increase, is the net saving. Other initiatives at the site (a new carrier contract) are excluded by measuring only cost centre CC-410 labour, agency and equipment lines.

Measure benefits with the method agreed at approval, and apply it consistently. Three adjustments are almost always needed. First, adjust the baseline for volume, because a growing business would have spent more in the do-nothing case; savings measured against an unadjusted baseline are understated in growth and overstated in decline. Second, adjust for input prices such as wage inflation, using the actual rates rather than the planning assumption. Third, isolate the project from other initiatives at the same site, either by measuring only the cost lines the project affects or by estimating the other initiatives' effect separately. Document each adjustment and its source so that the next review can repeat it.

Distinguish run-rate from cumulative benefits. A project can reach its run-rate (the annual saving once everything works as designed) and still be behind on cumulative benefits because it started late, as in the example. The approving body needs both views: the run-rate tells it whether the asset performs, the cumulative figure tells it whether the investment case still holds.

Unplanned benefits and disbenefits: Example: fewer manual-handling injuries (3 reportable incidents in FY2026, none in FY2029 to date); disbenefit: a single control-software supplier now holds a critical dependency (see REC-003).

6. Financial re-appraisal

6.1 Re-appraisal cash flows (valuation date 2027, 9.0%)

Notes on the example:

• Year 0: capex paid in FY2027 2,610 plus spare parts 120 = 2,730. Year 1 capex 400.
• Tax depreciation: actual capex 3,010 / 8 = 376.25 a year. Year 1 tax: 0.30 x (330 - 376.25) = -13.9, a tax credit used within the German tax group, so it adds to cash flow: 330 + 13.9 - 400 = -56.1. A full year of tax depreciation is kept in year 1 for comparability with the approved model, although the asset was available for use only from 2028-04-01 and actual German tax depreciation is time-apportioned (9/12 in 2028). The effect on NPV is small; disclose it rather than change the basis.
• Savings from FY2030 are re-forecast at 98% of the approved profile, rounded to whole EUR thousand, because the run-rate in FY2029 is about 2% below plan on a volume-adjusted basis.
• Year 8: disposal value 150 less tax 45 = 105, plus spare parts released 120 = 225. Net cash flow 751 - 112.4 + 225 = 863.6 (rounded from unrounded tax).
• Present values use unrounded factors and agree with the four-decimal factors to within 0.1.

6.2 Metrics: approved versus re-appraisal

Payback for the re-appraisal: cumulative undiscounted cash flow is -394.2 at the end of year 5, so 5 + 394.2 / 617.6 = 5.64. Discounted payback: 7 + 293.8 / 433.4 = 7.68.

6.3 NPV bridge

The capex steps include the change in tax depreciation that follows from the higher cost.

6.4 Forward-looking view

Sunk costs are irrelevant to decisions still open. Example: at the review date the remaining decisions are whether to add the second sortation loop (real option recorded in APR-2027-002 section 9.2) and whether to extend the software support contract. Volume is about 1.17 million orders a year (6% above FY2026), below the 1.5 million trigger, so the expansion is not yet proposed. The asset shows no impairment indicator: re-forecast cash flows comfortably exceed the carrying amount; the controller confirms this under the group impairment policy.

7. Variance causes

8. Accuracy of the original assumptions

Example checks: 210 / 2,800 = 7.5%; 20 / 100 = 20%; (330 - 520) / 520 = -36.5%; 10 / 680 = 1.5%.

Finding (Example): the appraisal tested cost and savings ranges well, but did not test timing. A one-quarter delay cost almost as much NPV (VAR-D, EUR 122.0k) as the whole capital overrun net of phasing (VAR-A plus VAR-B, EUR 133.4k). Future appraisals should include a delay scenario.

9. Lessons learned

9.1 Making lessons stick

Lessons only have value if they change what happens next time. Route each lesson to the document or process it should change: the appraisal template and its guidance, the capex request checklist, the project delivery method, or the benefits tracking routine. Record the change in that document's revision history with a reference to this review, so that a later reader can see why the rule exists.

Look across reviews as well as within them. Keep a simple register of the main variances from every completed review, by project type: cost overrun, delay and benefit shortfall as a percentage of the approved figure. After a few years the register gives evidence-based ranges for sensitivity analysis (for example, typical delay for projects with building works) and shows whether lessons from earlier reviews have reduced the errors they addressed. Present the register to the capital allocation committee once a year.

10. Recommendations and actions

11. Close-out

Sign-off: the sponsor confirms the project is closed and benefits are owned by [@benefit-owner]. Date: YYYY-MM-DD.

12. Glossary and references

12.1 Glossary

12.2 References

Standard and licence note

Based on common corporate practice for post-completion review of capital investments and on publicly available UK government guidance (HM Treasury Green Book, 2026 edition, and the Magenta Book, Crown copyright under the Open Government Licence v3.0), cited by reference only; no text from them is reproduced. Corvid Commerce Group, the project and every figure are fictional. This template is original text written for this collection. Check the current official versions of cited guidance before use. Template licence: MIT.