Most project dashboards can tell you the numbers. Few can tell you whether those numbers mean what everyone assumes they mean.
A KPI is only useful if it measures something that actually predicts or explains project performance β not just something that’s easy to collect and pleasant to report. The gap between those two things is where a lot of project control effort quietly goes to waste.
PERCENTAGE COMPLETE IS THE CLEAREST EXAMPLE OF THE PROBLEM
Percentage complete is one of the most widely used project metrics, and one of the least reliable on its own. It’s simple to calculate, easy to visualise, and satisfying to watch climb toward 100%.
But it rarely distinguishes between the easy 80% and the difficult 20%. Early progress on a construction project often moves quickly β mobilisation, groundworks, repetitive activities with well-understood risk. The remaining portion frequently concentrates exactly the elements that are hardest to control: interface coordination between trades, late design information, commissioning, approvals that depend on third parties, and any variation or claim still working its way through resolution.
A project reported as 80% complete can, in practice, still be carrying 50% or more of its real delivery risk. The number is accurate. What it implies β that the hard part is mostly behind you β often isn’t.
A GREEN STATUS IS A THRESHOLD, NOT A DESCRIPTION
RAG status reporting has the same structural weakness, for a different reason. Green, amber and red are usually defined by thresholds set once, early in the project, and rarely revisited.
A schedule variance of five days might have been comfortably green when the programme had six months of float. The same five-day variance, on the same threshold, can be genuinely dangerous once float has been consumed elsewhere. If nobody adjusts the threshold as the project’s risk profile changes, the colour keeps reporting the old context, not the current one.
This is why a green dashboard can sit directly next to a project quietly losing control β the status isn’t lying, it’s just answering a question that’s no longer the right one to ask.
A METRIC CAN BE ACCURATE AND STILL MEASURE THE WRONG THING
The same problem shows up in how many organisations track safety. Incidents reported is a common KPI β easy to define, easy to count, and genuinely important. But it measures outcomes that have already happened, not the conditions that make future incidents more or less likely.
Two projects can report an identical number of incidents while having very different underlying risk profiles β one with disciplined near-miss reporting and active hazard management, the other simply not yet unlucky. The KPI treats them the same, because it was never designed to see the difference.
The same pattern applies well beyond safety. A cost KPI that tracks committed spend against budget can look perfectly healthy while completely missing unresolved variations that haven’t yet been formally recognised. A quality KPI that counts inspections completed says nothing about whether the inspections found anything, or whether the findings were actually closed out.
WHY THIS MATTERS MORE THAN IT SEEMS
None of this means percentage complete, RAG status, or incident counts are useless. It means that treating any single metric as a complete picture β rather than one input requiring interpretation β creates a specific kind of blind spot: the project looks measured, controlled, and reported, while the thing that actually determines the outcome moves largely unobserved.
This is a quieter failure mode than having no data at all. A project with no KPIs at least knows it’s flying without instruments. A project with the wrong KPIs believes it has visibility it doesn’t actually have β which is often more dangerous, because it removes the instinct to look further.
THE TEST WORTH APPLYING BEFORE ADDING ANOTHER METRIC
Before adding a new KPI, or trusting an existing one, three questions are usually more useful than the metric itself:
What decision does this number actually inform? If nobody can point to a specific decision that would change based on the metric moving, it’s probably decoration, not control.
Who acts differently because of it? A KPI that everyone reviews but nobody owns tends to sit on a dashboard without triggering anything. Ownership β a named person who’s expected to respond when the number moves β is usually what separates a working control from a reporting artefact.
Does it explain a cause, or just describe an outcome that’s already happened? Lagging indicators (percentage complete, incidents to date, cost variance) confirm what already occurred. Leading indicators β unresolved RFIs, overdue approvals, productivity trends on critical activities, open interfaces β are what give management time to act before the outcome is locked in. A dashboard weighted entirely toward lagging indicators can look busy and still leave management perpetually reacting to news that’s already old.
THE PRINCIPLE
A dashboard full of accurate numbers can still fail at its one job: helping someone make a better decision, sooner.
Dashboards exist to compress information into something a decision-maker can absorb quickly. That compression is valuable β but only as long as what gets compressed away is noise, not the signal the decision actually depends on. What you choose to measure quietly defines what the project ends up managing. Everything left off the dashboard tends to receive less attention by default, whether or not that was the intention.
The right question isn’t “do we have enough KPIs?” It’s “does each one earn its place by informing a real decision, owned by a real person, early enough to matter?”
How ACC TRUST can support
ACC TRUST supports project owners, contractors and lenders in strengthening the connection between project reporting and actual project control, including:
- reviewing existing KPIs and dashboards against the decisions they’re meant to support;
- identifying leading indicators appropriate to the specific risks of a project, not just standard templates;
- assessing whether RAG thresholds still reflect current project risk, or reporting against assumptions set at baseline;
- structuring decision ownership so that KPI movement triggers action, not just observation;
- independent review of project reporting where the picture presented needs to be verified against underlying detail.
For independent support on project controls, reporting structures or contract governance:
Explore ACC TRUST services:
β https://acctrust.ro/en/services
Discuss a specific project:
β office@acctrust.ro
About ACC Trust Insights
ACC Trust Insights is the knowledge centre for Commercial & Contract Governance, Project Delivery and Risk Management in complex construction, infrastructure and energy projects.
Explore all articles:
β https://insights.acctrust.ro
ACC TRUST
Commercial & Contract Governance Advisory
Property Β· Infrastructure Β· Energy