A forklift safety program without KPIs is a program without a dashboard: the decisions run on memory and the results run on hope. The leading vs lagging post on this site sorts the measure families; this post is the concrete list, the metrics every warehouse should track, the targets that make them useful, and the review that turns the numbers into prevention.
The leading set
The leading KPIs are the process measures, and they are the ones the weekly review manages by. The near-miss rate, reports per shift, with the PSIF share broken out, is the precursor stream. The inspection completion rate, the share of daily inspections done on time, is the maintenance foundation. The training currency, the share of operators with valid certification, is the competence layer. The speed events, the share of zone crossings over the limit, is the behavior layer. And the at-risk behavior rate, from the observation log, is the BBS measure. The five read together tell the operation where the exposure is building.
The lagging set
The lagging KPIs are the outcome measures, reviewed monthly. The recordable injury rate per 100 operators, the severe injury count, the impact count from the telematics log, and the downtime hours per truck. The lagging set confirms the leading story: the precursor rate falls, and the injury and impact rates follow. The data on this site is the benchmark context, the national severe injury patterns that the operation’s numbers are measured against.
The operational set
The third family is the operational KPI that the fleet management runs on: the utilization rate, the maintenance overdue count, the cost per truck, and the fleet right-sizing. These are not safety KPIs by name, and they are safety KPIs in effect, because the right-sized, maintained fleet is the fleet with less pedestrian exposure and fewer failure events. The asset management post on this site covers the family.
The targets
The targets make the KPIs useful, and they should be set from the operation’s own baseline, not from a generic chart. Measure the current values for a quarter, then set the targets: the near-miss rate rising (reporting is improving), the PSIF share falling, the inspection completion at 100 percent, the training currency at 100 percent, the overdue maintenance at zero, and the speed compliance trending up. The zero incident process on this site is the target structure, and the Risk Scorer provides the exposure score that the targets move.
The review
The data sources
The KPIs are only as good as their data sources, and the source map is worth naming. The near-miss and the PSIF rates come from the reporting channel. The inspection completion comes from the maintenance software. The training currency comes from the certification records. The speed events come from the telematics log. And the at-risk behavior rate comes from the observation program. Each source has an owner, and the monthly review verifies the sources are feeding before it reads the numbers. The incident platform on this site consolidates the streams, and the Risk Scorer turns the consolidated data into the exposure score. The KPI review is the place where the sources and the numbers meet.
The KPIs live or die in the review: the weekly read of the leading set, the monthly read of all three families, with the owners and the fixes assigned, and the quarterly rescore that checks the targets against the changed operation. The monitoring review on this site is the cadence, and the incident platform is the record. The dashboard is only the tool; the review is the program, and the operation that reviews its KPIs weekly, with the data complete and the fixes closed, is the operation whose injury trend follows the numbers down. The KPI set is deliberately small, because the operation that tracks five numbers and reviews them weekly beats the operation that tracks fifty and reviews them annually, and the five on this page are the five that move the exposure. The five leading numbers and the three lagging ones are the whole scoreboard, and the operation that reads them on the fixed cadence, with the data sources verified, is the operation whose review meetings produce decisions instead of discussions.
