The safety scoreboard has two kinds of numbers, and the operation that reads only one is flying blind. The lagging indicators record what happened: the injuries, the serious events, and the incident rates that the data on this site tracks. The leading indicators predict what will happen: the near-miss rate, the PSIF rate, the inspection completion, and the speed events. This post sorts the two families and builds the scoreboard that uses both.
The lagging family
The lagging indicators are the outcome measures: the recordable injury rate, the lost-time cases, the severe injury count, and the incident frequency. They are the measures that the regulator and the insurer read, and they are the honest bottom line. Their limit is the statistical problem on this site: the serious events are rare, so the lagging numbers are zero for long stretches, and a zero that lasts for months looks like safety while the exposure builds underneath. The lagging measure cannot tell the operation what to fix this week, because it records what already happened.
The leading family
The leading indicators are the process measures: the near-miss and PSIF rate, the at-risk behavior rate, the inspection completion rate, the training currency, the maintenance overdue count, and the speed compliance. They are the measures that the zero incident process on this site manages by, because they change week to week and they point at the fix. The leading indicator is the precursor: the near miss that is reported today is the serious event that is prevented next quarter, and the rate is the leading edge of the outcome.
The connection
The two families connect through the PSIF framework: the leading indicators track the precursor stream, and the lagging indicators track the outcome stream, and the severe injury data on this site is the evidence that the connection is real, the mechanisms that produce the serious events appear in the near-miss log first. The operation that watches the leading numbers move and the lagging numbers follow is reading the connection correctly: the precursor rate falls, and the injury rate follows after. The operation that celebrates the lagging zero while the leading numbers worsen is reading the scoreboard backwards.
The scoreboard
The scoreboard uses both families, with the emphasis on the leading. The monthly review reads the KPIs on this site: the leading set, the near-miss rate, the inspection completion, the training currency, the maintenance overdue, and the speed events, plus the lagging set, the injury counts. The Risk Scorer adds the exposure score that combines both, the score falls when the leading measures improve. The weekly review on this site works from the leading numbers, because the weekly fix has to be made before the outcome, and the monthly review reads the lagging confirmation.
The honest frame
The review cadence that uses both
The two indicator families need different review cadences, and the cadence is worth naming. The leading indicators are read weekly, because the near-miss rate, the inspection completion, and the speed events change week to week and point at the immediate fix. The lagging indicators are read monthly and quarterly, because the injury counts are too rare and too slow for the weekly read. The Risk Scorer is the bridge that is rescored quarterly, combining the leading inputs into the exposure score that the lagging outcomes are expected to follow. The monitoring review on this site is the weekly slot, and the zero incident process is the monthly structure. The operation that reads the leading numbers weekly and the lagging numbers monthly is the operation that manages the future and confirms the past.
The honest frame is that the lagging indicators are the truth and the leading indicators are the control. The operation cannot manage the injury rate directly; it can manage the precursor rate, the behavior, the conditions, and the systems, and the injury rate follows. The zero incident process on this site is the structure that runs the leading measures with the lagging goal, and the data on this site is the context that keeps both honest. Lead with the leading, confirm with the lagging, and the scoreboard shows the operation where it is going, not just where it has been.
