Near-miss reporting collapse and the £150 zero-harm bonus

What the numbers probably mean

• The fall from about 64 to 11 cards per quarter began immediately after the supervisor was warned. That timing, across all three sites, is far more consistent with a change in reporting behaviour than with an 80% improvement in conditions. A warning that follows a report teaches the workforce that reporting predicts punishment; rational workers then stop reporting.
• The cards still arriving are mostly housekeeping, which suggests people now raise only the reports they judge 'safe'. Serious near misses are likely being absorbed into the informal system, or not discussed at all.
• LTIFR rising from 1.7 to 3.2 while recordable injuries are 'broadly unchanged' needs care with small numbers. Across roughly 550 people, one or two extra lost-time cases can move the rate sharply. This is a signal to investigate, not proof that risk has doubled. Check severity mix, days lost per case and a rolling 12-month rate rather than a single quarter.

Why the measures are weak

• LTIFR is a lagging indicator, heavily influenced by reporting culture and by whether injured people take time off rather than by the standard of control. It reports harm already done and says nothing about control failures that have not yet produced an injury.
• A £150 bonus payable only when there is no lost-time injury creates a direct financial incentive to keep injuries off the record and to press injured colleagues to return early. It rewards the lucky depot rather than the safe one. The measure can improve while the outcome worsens.
• 'Zero harm' delivered as a slogan with a cash trigger reframes reporting as disloyalty. It is a target, not a strategy.

Separate the two questions

The supervisor's failure to follow the segregation procedure is a conduct issue and should be managed as one, proportionately, on its own merits. Whether he should be disciplined for reporting is a different question, and the answer is no. What the workforce observed was: report → investigation → written warning. The organisation cannot now argue that reporting is welcome without evidence that this has changed.

Practical steps

1. A senior leader statement, given in person at each depot, that reporting a near miss or hazard will never in itself trigger disciplinary action, and that the warning will not be used as an example. Consider whether the warning should be reviewed; at minimum, explain the distinction publicly.
2. Remove the lost-time bonus. Replace it with something the workforce cannot manipulate by silence, such as closure of reported hazards within an agreed timescale.
3. Close the loop visibly: a weekly 'you said, we did' board at each gatehouse, with the reporter's name withheld.
4. Include agency workers formally. They are 130 of about 550 people and are commonly left out of schemes, inductions and feedback; excluded groups report least.
5. Make reporting easy and blameless: a short card or app, an anonymous route, translated forms where needed, and no requirement to name yourself.
6. Classify reports by potential consequence, not just by outcome, so a serious near miss gets a serious response even when nothing was damaged.
7. Train supervisors in just-culture conversations: distinguish human error, at-risk behaviour and reckless conduct, and respond proportionately to each.

Reasoning

Behaviour follows consequences. If reporting produces an investigation, blame and a warning, silence is the rational response. Restoring reporting requires credible evidence that the consequence has changed, not another poster or another prize draw.

Design principles

• Few enough to be read and challenged; each with an owner, a written definition and a data source.
• A mix of leading and lagging, and of hard data and qualitative evidence.
• Nothing tied to individual pay, because any measure that pays for a number will be gamed.
• Small numbers handled with rolling 12-month rates or funnel plots so that normal variation is not read as a trend.

Suggested set

• Leading: percentage of planned workplace inspections completed, and percentage of actions closed by the agreed date.
• Leading: near-miss reports per 100 workers, with the proportion rated 'serious potential'. The aim is a healthy or rising reporting rate with a falling severity profile, not a falling rate.
• Leading: percentage of racking, forklift and pedestrian-segregation checks passed; number of forklift-pedestrian conflict events.
• Leading: competence currency - inductions, refresher training and agency worker inclusion.
• Lagging: lost-time injury frequency rate and severity rate (days lost per 1,000 hours worked), rolling 12 months.
• Lagging: reportable incidents and dangerous occurrences.
• Qualitative: a short anonymous safety climate survey twice a year, plus a sample of worker interviews carried out by someone independent of the line.

Using them at the board

The set answers different questions: are controls in place (leading), are they working (assurance checks), and what harm occurred (lagging). A falling LTIFR alongside falling near-miss reporting and falling inspection completion is a warning, not a success. Management review should interrogate the story the numbers tell - why has reporting fallen, what is happening on the floor - rather than simply noting the figures. Perverse incentives arise when one number becomes a target, so the board should see them together and ask for the narrative behind each movement.