Project KPIs That Earn Their Place: Good Metrics vs. Bad Ones
Every troubled project I have reviewed had a dashboard. Usually a colourful one, updated faithfully, full of numbers nobody acted on. The problem is rarely a lack of metrics; it is a surplus of the wrong ones. A good key performance indicator changes a decision. A bad one decorates a report. With teams now spread across home offices and hybrid schedules, where you cannot just walk the floor to sense how things are going, the gap between those two kinds of numbers has only grown more expensive.
What a bad KPI looks like
Bad KPIs feel productive because they are easy to collect and almost always look fine. They measure motion instead of outcome.
Percent complete with no definition of done — a number that drifts to 90% and stays there for months.
Hours logged, treated as if effort were the same as progress.
Number of meetings held or emails sent, which reward activity and punish nobody.
A single overall 'health' colour with no underlying measure, so green means whoever owns the report feels optimistic.
Vanity totals like 'tasks closed' that say nothing about whether the right tasks were closed.
What a good KPI looks like
A good KPI is tied to an objective, has a clear target, and points to an action when it moves. You should be able to finish the sentence: 'If this number goes the wrong way, we will ___.' If you cannot, it is not a KPI, it is trivia.
Schedule Performance Index, not raw percent done. Comparing earned value to planned value tells you whether you are ahead or behind in a way a self-reported percentage never will.
Cost variance against budget, read as a trend. One month means little; three months sloping the wrong way is a decision.
Forecast at completion, refreshed each period. Where will we actually land on time and cost if nothing changes? That is the number sponsors need.
Open critical risks and aging of issues. A risk that has sat unaddressed for six weeks is telling you something a status colour hides.
Defect or rework rate. Speed means nothing if you are redoing the work; quality leakage shows up here before it shows up in the budget.
Choosing the few that matter
You do not need fifteen indicators. You need the handful that connect to what this specific project must deliver. Start from the objectives, ask what would have to be true for each to succeed, and let that suggest the measure. Pick a small set, define each one precisely — including what 'done' and 'on target' mean — and agree who owns the number and how often it is refreshed. Make the data easy to gather honestly; if a KPI is painful to update, people will guess, and a guessed metric is worse than none. Review them in a rhythm that matches the project, and retire any indicator that has not changed a single decision in three reporting cycles. A short, trusted set beats a sprawling dashboard nobody believes.
The test never changes: a metric earns its place on the page only if someone would do something different when it moves. Everything else is noise dressed up as insight.
If your reporting feels busy but your decisions still feel blind, it may be time to rebuild the few measures that count — something XNM's program & project delivery advisory helps clients do every day.