A near miss is an event that could have caused injury or damage but did not, often by a matter of centimeters or seconds. It is also the most under-used data in workplace safety. In a forklift operation, near misses happen daily, and they share the same root causes as the serious injuries in the record books. This post defines near misses with concrete examples, explains why they are systematically under-reported, and lays out a reporting program that actually generates data. It sells none of these systems.
What counts as a near miss
A near miss is defined by potential, not outcome. The classification questions are simple: did an unplanned event occur, could it have caused injury or damage, and did it not, this time. In a forklift operation, the classic examples are: a pedestrian steps into an aisle as a truck rounds a corner and the operator stops centimeters short; a load shifts on the forks during a turn and settles back; a truck backs within inches of a rack; a worker reaches between a truck and a fixed object as the truck moves. None of these produced an injury. All of them carried the potential, which is what makes them signals rather than noise.
Why near misses matter more than they look
The value of near misses comes from their frequency and their shared root cause with serious events. The severe-injury patterns in this site’s dataset, 5,154 forklift-involved reports across 2015-2025, are dominated by struck-by and caught-between events, and every one of those injuries was a near miss seconds earlier. The accident statistics guide lays out that distribution. If a safety program can see the near-miss versions of those mechanisms, it can act on the causes while the cost is still zero.
The under-reporting problem
The structural fact about near misses is that they are under-reported. Most organizations rely on voluntary reporting, and the fraction that gets reported is small. The reasons are human: the operator who nearly caused an incident does not want to admit it, production pressure discourages stopping, and a history of blame teaches workers to stay quiet. The fix has two parts. The cultural part is a blame-free policy with a visible response to every report, which the safety rules guide structures. The technical part is automatic logging: proximity detection and telematics systems record alarm events with time, location, and identities, whether anyone reports them or not. The automatic log changes the denominator from a handful of voluntary reports to the actual event stream.
Near-miss categories in a warehouse
Warehouse near misses sort into a handful of categories, and knowing the categories makes the log useful. The struck-by category covers a pedestrian entering the path of a moving truck, which is the largest family and the one pedestrian detection targets. The caught-between category covers a worker pinned between a truck and a rack or a dock, addressed by collision avoidance. The load category covers shifting, falling, or overheight loads. The equipment category covers the inspection defects, a failing brake or a cracked tire, that the daily inspection exists to catch. The behavior category covers speeding, phone use, and the other unsafe acts that appear in observation. Tagging each report with its category turns the log into a dashboard, and the dashboard is what a weekly review can actually act on.
The severity tag is the second field worth adding, because frequency and potential do not move together. A category that produces ten low-potential events a week and a category that produces one high-potential event a month need different responses: the first needs housekeeping, the second needs engineering. The PSIF framework on this site is built on exactly that distinction.
Building the reporting program
A working near-miss program has four elements: a simple reporting channel that takes less than a minute, a blame-free policy stated in writing, a response loop where every report gets a visible outcome, and recognition for reporters. The forklift inspection checklist catches the equipment-side conditions, while the report log catches the behavior-side events. Review the combined log weekly, look for hotspots, shift-change spikes, and repeat offenders, then intervene and measure whether the rate falls.
The starting point is to score the operation and build the reporting around it. Run the Risk Scorer to see where your exposure sits, compare the control families on the systems index, and when you are ready for vendor numbers, send your scope through the quote form. The methodology page documents how the scorer derives its inputs.
