Warehouse metrics
A record of what happened,not a forecast of what will.
Safety incidents is the count of recorded incidents and near-misses from your own reporting process, set against the conditions the twin can measure on the pick face that produced them — golden-zone share, handling weight at height, travel and repetition across a route. WareBee doesn't predict who gets hurt next; it reads what already happened alongside what's measurable about the conditions around it, plainly and without a forecast attached.
See the fatigue and injury risk challenge
What it is
What happened, next to what the pick face was asking.
On WareBee's digital twin, safety incidents is a count: recorded incidents and near-misses drawn from your own reporting process, set alongside the conditions the twin can actually measure on the pick face involved — golden-zone share for the SKUs on that route, the handling weight assigned to the location, and the travel and repetition the route carried across the shift. WareBee doesn't generate the incident record itself and doesn't predict which associate is at risk next; it reads what was reported next to what's measurable about where and how the work happened, and nothing further.
The figure moves for two separate reasons, and they shouldn't be read as one. How much gets reported changes with how consistently near-misses get logged at all, independent of whether conditions on the floor have changed. And separately, the conditions the twin can measure change with slotting — a fast mover drifting outside the golden zone, a heavy case assigned without checking the height band, a route accumulating more bend, reach and carry than the ones around it — whether or not any incident has yet been recorded against them.
A raw incident count says little without knowing how much handling and travel the operation actually generates — a site running heavier pallets across a larger footprint carries more physical exposure by design than a smaller, lighter one, independent of how well either is run. WareBee reads your count, and the conditions logged alongside it, against a peer set matched on order profile, SKU count and footprint, so the comparison starts from operations carrying a genuinely similar physical load.
What moves this number
The conditions the twin can measure on the routes involved.
Each one is read from your digital twin from data you already generate — a set of conditions to check, not a verdict on what caused any single incident.
Fast movers drifting outside the golden zone
The SKUs picked most often sit outside the height band where reaching and stooping are minimised.
Golden-zone share tracks each SKU's placement in its height band against how often it's actually picked, not a fixed slot chosen once and left. When a fast mover drifts outside that band, the golden-zone share for that route drops before anyone connects it to the extra reach or stoop it adds to every pick.
Weight assigned without checking the height band
Heavy items land wherever there's space, not wherever lifting them is safest.
Mass and handling class are checked against a location's height band before a slot is confirmed, where that check is configured to run. A heavy case assigned to a top shelf without it shows up as a failed check on the twin — a condition worth reviewing, not a verdict on whether it caused anything.
Travel and repetition building unevenly across the shift
Bend, reach and carry distance accumulate on a route, and the load doesn't always spread evenly across the team.
Bend, reach and carry distance are tallied per route, not just pick count, so two routes with the same number of lines can carry very different physical loads. Where that load keeps landing on the same associates rather than being spread across the shift, the twin shows the pattern independent of whether an incident has been logged against it yet.
Conditions reviewed only after something's already happened
Placement risk usually surfaces in an incident report, not in a check that ran before the shift started.
Without a continuous check, a placement condition is often only reviewed after an incident report names it, by which point the layout may already have changed. Logging every placement decision against the rule it was checked against — weight limit, reach height — gives a reviewer a record that already exists, rather than one reconstructed from memory afterwards.
Read the record next to what the twin can measure.
WareBee doesn't turn an incident count into a prediction, and it isn't a substitute for a health-and-safety professional or a formal risk assessment. What it does is place your recorded incidents and near-misses next to golden-zone share, handling weight and travel-and-repetition load for the routes involved, so a pattern in the conditions — if there is one — is visible in the same view as the record itself, rather than argued from memory.
Where a condition looks worth changing, the re-slot is simulated against your real orders before a pallet moves, so the travel and handling-load effect is known first, not discovered on the floor afterwards. Once your team approves a change, it reaches the WMS as the moves and location assignments it already understands.
- Recorded incidents and near-misses set against golden-zone share, weight and travel load
- No prediction of who is at risk — a record, read next to measurable conditions
- Re-slotting simulated against real orders before a pallet moves
Questions
Safety incidents, read plainly.
What exactly does the safety incidents measure show?
It's a count of recorded incidents and near-misses from your own reporting process, set alongside the conditions the twin can measure for the routes involved — golden-zone share for the SKUs picked, the handling weight assigned to the location, and the travel and repetition the route carried. WareBee doesn't generate the incident record itself; it reads what was reported next to what's measurable about where and how that work happened.
Does this measure predict who's likely to get hurt?
No, and WareBee doesn't claim otherwise. It doesn't run a RULA, REBA or NIOSH-style assessment and it isn't a substitute for a health-and-safety professional's judgement, and it does not predict which associate is at risk or forecast a future injury. An incident count is a record of what already happened; golden-zone share, handling weight and travel load are conditions the twin can measure today — reading them together shows a pattern if one exists, not a prediction of what comes next.
Where does the incident and near-miss data come from?
From your own incident-reporting process — however your operation already logs incidents and near-misses, whether that's a WMS module, a dedicated H&S system, or a manual log brought in as a CSV export. WareBee doesn't originate this data or judge its completeness; it takes the record as reported and reads it alongside the golden-zone, weight and travel conditions the digital twin already measures for the same routes and locations.
Why read incidents against golden-zone share and travel load instead of a single safety score?
Because those are the conditions on the pick face the twin can actually measure, rather than a single score that would compress reaching, weight and repetition into one number and lose what each was telling you separately. A route that's heavy but well within the golden zone is a different situation from one that's light but constantly reaching above shoulder height, and collapsing the two into a score would hide which applies.
Doesn't a published injury rate make a more meaningful comparison than a peer set?
Because a published injury rate would have to average over every kind of warehouse there is, and physical exposure doesn't sit still across that range — a site moving heavier pallets across a larger footprint carries more handling and travel by design than a smaller, lighter one, independent of how well either is run. WareBee's peer set holds order profile, SKU count and footprint constant instead, so the comparison starts from operations carrying a genuinely similar physical load.
Once a flagged condition is fixed, does that change reach the WMS?
Yes. Once a re-slotting plan addressing a flagged condition — a fast mover moved into the golden zone, a heavy SKU reassigned to a safer height — is approved, it goes out as the location assignments and putaway moves the WMS already runs on. Pick and location data flow back afterwards, so the next reading reflects what's genuinely sitting where, not what the plan intended.