Case study · Human error
SENTINELLE programme: 47% fewer “human error” deviations in 6 months
Repeated retraining with no lasting effect: this site changed its approach. A human error expert per area, trained to become autonomous, and short shop-floor rituals to handle weak signals before they turn into errors.
The results
The starting point
- A packaging unit of about 200 people, in 2 shifts, with very manual operations.
- 2022: 408 deviations, 108 of them due to human error, i.e. 26%.
- Repeated retraining with no lasting effect.
What we did together
- AssessA flash assessment of the site, then an introduction for the teams.
- Appoint expertsA “human error expert” per area (packaging, warehouse, PSL / PPL), coached to become autonomous.
- Adapt the toolsA tool to help analyse human errors, adapted to the site.
- Put in place 4 ritualsWeak signal reporting, task observations, “15 minutes for Quality” and a short-term action plan.
- May → Oct. 2023Coach for 6 monthsShop-floor coaching, until the experts are autonomous.
Errors fall in the coached areas
“Human error” deviations opened per month, on average, in 2022 and then in 2023.
Weak signals handled before they become errors
set to different ranges: placed at either end, with colour coding.
awaiting inspection: the waiting area was redesigned with the shop floor.
of labels and leaflets: criteria defined and a double check.
held up at customs: invoicing software, so the patient receives their treatment.
Our takeaways
Task observations created strong moments of awareness, and “15 minutes for Quality” was well adopted. The experts improved from 71% to 79–100% on their observation grid, against a target of 75%.
The approaches used
Our advice if you’re getting started
- Keep the rituals alive between the coach’s visits: that’s where everything happens.
- Rely on internal experts, as close as possible to the shop floor, rather than on retraining.
Do human errors keep coming back despite training?
Let’s talk about your situation: we’ll show you how to handle weak signals before they become deviations.
