Guide · 6 min read
Measuring whether your training is working, with three numbers
One: coverage per station
How many people are currently signed off on each station, counting only sign-offs that have not expired. This is the operational number and it is the one that predicts a bad night.
Read it by looking for the thinnest column rather than the average. A crew averaging four people per station can still have one station covered by two, and that station is one call-out from being a problem.
Two: score trend per station
Whether average scores on a station are rising, flat, or falling over months. Rising means the training is landing. Flat at a high number is fine. Falling is the interesting case.
A falling trend on one station usually means something changed there: new equipment, a new build, a menu change nobody retrained for, or the trainer who used to run it left. It is a prompt to look rather than a verdict.
Three: the most-missed item
Which specific checklist item gets missed most often across the whole crew. This is the highest-value number in the category and almost nobody computes it.
When one item is missed by most people, the odds that twenty individuals independently have the same weakness are low. Far more likely: the material is unclear, the equipment makes it awkward, the standard was never actually communicated, or the item is written in a way two trainers read differently.
It is the cheapest fix available to an operator, because changing one line of training material or one piece of station setup improves everybody at once. It is also completely invisible without scoring against a fixed checklist.
What not to measure
Training hours delivered. It is an input, it is easy to count, and it correlates with nothing. An operator can double it and change nothing about whether the floor is covered.
Completion percentage, on its own, has the same problem: it tells you material was consumed, not that anybody can do the job.
- Coverage per station, counting only unexpired sign-offs
- Score trend per station, over months rather than weeks
- Most-missed checklist item, across the crew
- Optionally: time from hire to first sign-off, as a check on your own scheduling
You need a fixed checklist for any of this
All three numbers depend on everybody being scored against the same items. Free-text evaluations cannot produce any of them, because there is nothing to aggregate.
That is the practical argument for a structured checklist over a paragraph, and it is worth being clear that it is the only real reason. A paragraph is often a richer description of one person. It just cannot answer a question about twenty.
Common questions
What are good training KPIs for a restaurant?
Coverage per station, score trend per station, and the most-missed checklist item. Those three answer whether you are staffed, whether performance is moving, and what to change. Training hours delivered is an input and should not be on the list.
How do you know if training is actually working?
Coverage on your thinnest station going up, and scores on a station trending up over months. If both are flat while training hours are being spent, the training is happening and not landing, which is a content or a trainer question rather than a volume one.
What does it mean when everyone misses the same item?
Nearly always that the problem is not the people. Check whether the item is worded ambiguously, whether the equipment makes it awkward, whether the standard was ever explicitly taught, and whether two trainers are interpreting it differently.
How often should you look at these numbers?
Coverage weekly, because it is operational and changes with the roster. Trend and most-missed monthly, because they need enough evaluations behind them to mean anything and will look like noise if you check them too often.
The record-keeping half, handled
CrewSharp holds the checklists, the sign-offs and the re-check dates, and keeps station coverage, score trends and the most-missed item current on their own.
30 days free. No card to start. Cancel in one click.
Keep reading
How to Train a New Server (Step by Step)
The order to teach it in, what to hold back until week two, and when to hand over a real section.
How to Evaluate Restaurant Staff (Without Guessing)
Scoring people so two managers reach the same answer, and what makes an evaluation actionable rather than decorative.
Who Should Train Your Restaurant Staff
Why your fastest cook may be the wrong trainer, and the one-to-one rule that decides it.
Free station training checklists
The checklists themselves, one per station, free to copy.
Compare the options
Spreadsheet, LMS, paper, and the tools you are probably weighing.