Key takeaways
A key performance indicator is a number tied to a decision. If a KPI moves and nobody does anything differently, it is a statistic, not a KPI. The test is simple: for every metric on the board, name the action a bad reading triggers. The ones with no answer are clutter.
This matters because the instinct is to measure everything. Modern systems can capture hundreds of signals, so dashboards fill up and the few numbers that should drive the shift get lost in the noise.
Lagging KPIs report the past: OEE last week, downtime yesterday, yield on the last run. They are essential for knowing where you stand, but you cannot change them.
Leading KPIs predict the near future: preventive maintenance compliance, training completion, time-in-yellow on operating limits. A slipping leading indicator is a warning you can still act on. A board with only lagging numbers is a rear-view mirror. The preventive maintenance schedule is where most leading maintenance KPIs come from.
Vanity metrics look impressive and change nothing. Total units produced all-time, a rising machine count, a dashboard with forty gauges: none of them tell an operator what to do at 2 p.m. on a Tuesday. Cut any metric that exists to look good in a report rather than to drive a decision on the floor.
Fabrico calculates the core manufacturing KPIs from live, automatically captured data rather than manual logs, so OEE, downtime, and yield rest on real events instead of guessed reason codes. Because OEE and CMMS share one platform, maintenance KPIs like MTBF and MTTR sit next to production KPIs in the same view, and a bad number can become a work order without leaving the dashboard. Fabrico is built and hosted in the EU with data residency in mind and is ISO 27001 certified. To see your KPI board on real data, book a demo.
Teams putting this into practice often review our roundup of the best production monitoring systems.
Fewer than most do. A core set of around six that people actually act on beats a wall of gauges nobody reads. Add a metric only when you can name the decision it drives.
For equipment-driven plants, OEE is the usual headline because it folds availability, performance, and quality into one number. But it is a starting point for investigation, not an answer on its own; the value is in the losses it points to.
Lagging KPIs measure outcomes that already happened (last week's OEE). Leading KPIs measure inputs that predict future outcomes (PM compliance). You need both, but a board with only lagging numbers leaves you reacting instead of preventing.
Two reasons: too many metrics, so none drive action, and untrustworthy data underneath, so people stop believing the numbers. Keep the set small and make sure the inputs are captured accurately, not guessed.