1. Why percent complete misleads
Counting signed ITRs against total ITRs gives a number that rises smoothly and then stalls at 95% for a quarter. It misleads for three reasons: records are not equal in effort, records are not equal in risk, and a subsystem at 95% is not 95% handed over — it is not handed over at all.
Completions is a gated process. Value is realised in steps when a subsystem passes a milestone, not continuously as records accumulate. Any measure that ignores gating will overstate progress right up until it is obviously wrong.
2. The metric set worth reporting
| Metric | What it tells you | Failure mode if ignored |
|---|---|---|
| Subsystems by gate status | How much of the plant has actually crossed MC, RFC and RFSU. | Activity looks healthy while zero subsystems close. |
| Weighted ITR completion | Progress adjusted for record effort and criticality. | Easy records inflate the curve. |
| Prerequisite-clear backlog | How much work can start tomorrow. | Crews sent to work that cannot be executed. |
| Punch raised vs closed rate | Whether the deficiency population is converging. | Punch grows silently until it blocks the gate. |
| Category A ageing | Blocking items sitting unowned. | Certificates fail at the meeting, not before. |
| Re-test rate | Quality of first-time execution. | Rework consumes float invisibly. |
| Signature turnaround | Time between execution and acceptance. | Work is done but not credited; gates slip on admin. |
Put a number on it
What is this costing your project right now?
Estimate the hours and cost lost to transcription, register consolidation and dossier assembly — with editable assumptions and the maths shown.
3. Weighting records honestly
A pressure test on a large piping subsystem and a nameplate check are both one ITR. Weighting corrects for this, and there are three defensible bases:
- Manhour weighting — each checksheet type carries an estimated duration. Most accurate, most effort to maintain.
- Discipline weighting — a coarse factor per discipline and record class. Cheap, adequate for reporting at project level.
- Criticality weighting — records on safety-critical elements carry more weight, so the curve reflects risk retired rather than paper produced.
Pick one, agree it with the client before the first report, and do not change it mid-project. A weighting change mid-stream makes every historical curve non-comparable, which is worse than a mediocre weighting applied consistently.
4. Reading the S-curve properly
The commissioning S-curve is not one curve. It is at least three, and their spacing carries the real information:
- Executed — records physically completed in the field.
- Accepted — records signed off by the receiving party.
- Certified — subsystems that have crossed their gate.
A widening gap between executed and accepted means the signature chain is the bottleneck — an authority or availability problem, not a field problem. A widening gap between accepted and certified means punch or prerequisites are blocking gates. Reporting a single blended curve hides both diagnoses.
5. Forecasting RFSU dates
A credible subsystem forecast is built from four inputs, not from extrapolating a curve: remaining prerequisite-clear work, remaining blocked work with its blockers named, open Category A punch with owners, and the availability of the resources qualified to execute and sign.
The practical test of a forecast is whether it changes behaviour. If the report tells you a subsystem is at 92% and the recovery plan is 'push harder', the metric is not doing its job. If it tells you three loops are blocked by an uncalibrated transmitter and one Category A punch is nine days unowned, the plan writes itself.
That is the level commissioning management is designed to report at: not a percentage, but the specific named obstacle between a subsystem and its next gate.
6. Cadence and audience
| Audience | Cadence | What they need |
|---|---|---|
| Field supervision | Daily | Prerequisite-clear work list, blockers, re-test scope. |
| Completions management | Weekly | Gate status by subsystem, punch ageing, signature turnaround. |
| Project leadership | Weekly / monthly | Certified subsystems against plan, forecast RFSU dates, top blockers. |
| Client and operator | Monthly | Certificate status, carried punch, dossier readiness. |
One dataset, four views. When each audience gets its numbers from a different spreadsheet, the meeting is spent reconciling reports rather than resolving blockers.
Check your own process
Score your handover readiness in eight questions
A structured self-assessment of register control, evidence capture, sign-off and gate logic — with a specific recommendation for every gap.
