Project delivery
Benefits realisation plan
Benefit, measure, baseline, target, where the measurement comes from, the owner after handover, and leading indicators watched during delivery rather than after it.
What it is
An output is what the project produces and a benefit is the change that follows from somebody using it. A reconciliation engine is an output and it is finished when it works. Eleven staff days a week becoming two is the benefit, and it arrives only if people stop doing the old process.
That gap is why the owner column is the one that matters and the one usually left blank. Benefits arrive after the project ends, the project manager has moved on, and accountability has to sit with somebody still in place who controls the operational behaviour the benefit depends on.
Leading indicators are the other half. Waiting until a benefit is due to find out whether it is coming is a post mortem. Something observable during delivery predicts each benefit, and a bad reading then is recoverable in a way that a missed benefit measured a year after closure is not.
When to use it
- A previous project delivered exactly what was specified and changed nothing.
- A project is closing and nobody in the business has been named to own the outcome.
- A sponsor wants to know whether last year's investment worked and there is no baseline.
- The business case has targets and nobody has said where the measurement will come from.
What is in it
- The benefits
- An output is not a benefit
- Why the owner column is the one that matters
- Leading indicators during delivery
- When the benefit does not arrive
- Review points
The template
Project. Settlement rebuild
Business case reference. BC-2026-014
Plan owner. The person accountable after the project closes
The benefits
| Benefit | Measure | Baseline | Target | Measured from | Owner after handover | First measurement |
|---|---|---|---|---|---|---|
| Reconciliation effort falls | Staff days per week on manual matching | 11 | 2 | Operations timesheet codes 4412 and 4418 | Operations manager, North | Three months after go live |
| Settlement breaks fall | Breaks per month | 4 | 1 | Break register | Operations manager, North | Three months after go live |
| Volume absorbed without headcount | Instructions per full time employee | 4200 | 6000 | Settlement volume report over establishment | Head of operations | Twelve months after go live |
| Regulatory reporting exposure falls | Reportable incidents per year | 2 | 0 | Regulatory incident log | Regulatory affairs | Twelve months after go live |
An output is not a benefit
The project delivers a reconciliation engine. That is an output, and it is finished when it works. Nobody is better off yet.
The benefit is eleven staff days a week becoming two, and it arrives only if people stop doing the manual process, which depends on training, on trust in the new matching, and on the old tools being switched off. The project can build the engine and still produce none of the benefit, and that is the ordinary case rather than a rare failure.
Why the owner column is the one that matters
Benefits arrive after the project ends, and the project manager has moved on. The register is usually blank in this column, which is the whole problem.
The owner has to be somebody who is still in place when the benefit lands, who controls the operational behaviour it depends on, and who is measured on it. A project manager cannot direct operational adoption, does not control the call centre or the settlement team, and keeping a temporary role open indefinitely is not a substitute for accountability sitting in the business.
Fill this column before closure. A plan handed over with no named owner has been filed rather than handed over.
Leading indicators during delivery
Waiting until the benefit is due to find out whether it is coming is too late. Each benefit has something observable earlier that predicts it.
| Benefit | Leading indicator | Watch from | What a bad reading means |
|---|---|---|---|
| Reconciliation effort falls | Proportion of breaks routed through the new queue rather than the spreadsheet | First region live | The old process is still running alongside, so no effort is being removed |
| Settlement breaks fall | Match rate on the trial migration data | Trial migration | The rules do not cover the real data, and the break rate will not move |
| Volume absorbed | Time to resolve one break | First region live | Handling is not faster, so more volume will need more people |
| Reporting exposure falls | Breaks resolved inside the reporting window | First region live | The exposure is unchanged whatever the total break count does |
A bad reading here is recoverable. A missed benefit measured a year after closure is a post mortem.
When the benefit does not arrive
The honest positions are that the output was not adopted, the output works and the benefit was overstated, or something outside the project moved.
Adoption is the common one and it is addressable, usually by reinforcement rather than more training. An overstated case is worth recording plainly, because the estimating habit that produced it will produce the next one too. External change is nobody's fault and still needs saying, so the organisation knows what it actually got.
None of the three is a reason to stop measuring. A benefit nobody checks is a benefit nobody can claim, and the organisation loses the ability to tell which of its investments work.
Review points
Fix the dates now rather than intending to review. Each benefit gets a first measurement date from the table above and a review at least annually until it is realised or written off.
The plan owner runs those reviews. Where a benefit has landed, record it and close the row. Where it has not, the review decides whether more work would bring it, whether the target was wrong, or whether to stop and say so.
Questions people ask
- Why can the project manager not own the benefit?
- Because they cannot direct operational adoption, do not control the receiving team, and the role ends. Keeping a temporary role open indefinitely is not a substitute for accountability sitting in the business.
- What if the benefit does not arrive?
- The honest positions are that the output was not adopted, that the case was overstated, or that something outside moved. Adoption is the common one and usually needs reinforcement rather than more training. An overstated case is worth recording plainly, because the habit that produced it will produce the next one.
- How long should benefits be tracked?
- Until they are realised or written off, reviewed at least annually. A benefit nobody checks is one nobody can claim, and the organisation loses the ability to tell which investments work.