A digital logbook proves nothing on its own. It proves control only when the person holding the tablet understands what a reading means, knows the limit it has to clear, and knows exactly what to do the moment it doesn’t. Train people to tap the screen and you get tidy dashboards. Train them to read the evidence and you get records that hold up when an inspector, an insurer, or a solicitor starts asking questions.
That gap — between using the app and understanding the evidence — is where most rollouts quietly come undone. The vendor runs a ninety-minute walkthrough of the menus and calls it training. Useful, but it covers the easy half. It does nothing about the flushing task marked complete by someone who ran the tap for four seconds, or the hot-water temperature logged as “in range” by an operative nobody told the range to.
Get the system right before you train anyone
Training on a half-built system teaches bad habits you then have to unpick. Before a single staff session, three things need to be true.
- The tasks, assets, control limits and escalation routes are actually configured in the software, and they match your current written scheme and risk assessment. If the app calls a reading “in range” when your scheme says otherwise, the records are worse than useless.
- You have decided who does what: who records routine temperatures and flushes, who reviews the data, who owns the exceptions. L8 expects the people carrying out control measures to be trained and competent for the task, and that role map is what training delivers against [1].
- Every user has their own login. Shared accounts destroy the one thing a digital logbook is meant to give you — a clean audit trail of who did what, and when.
If you are still mid-migration from paper, settle the parallel-running plan first; Paper vs digital logbooks: making the switch covers the switch itself.
A rollout that actually sticks
Treat the training as a sequence, not a single event. Each stage has a clear point at which you can say it is done.
- Map each role to the records it owns. Don’t train everyone on everything. A caretaker who logs weekly outlet temperatures needs a different session from the responsible person who signs off the month. Done when: every role has a named task list and you can say, without checking, who logs the cold-water temperatures on each floor.
- Train on real assets, not a demo site. Walk operatives through their own outlets, their own calorifier, their own sentinel taps. Muscle memory built on dummy data evaporates on day one. Done when: each person can complete a genuine task on a genuine asset and say, in a sentence, why that check exists.
- Drill the out-of-range case hardest. This is the part generic training skips and the part that matters most. What does an operative record when a hot tap reads low? Who do they tell? What do they do before they walk away? Done when: every operative can demonstrate a failed reading end to end — recorded, escalated to a named person, and a next action logged, not just a red flag left on a screen.
- Decide what happens with no signal. Plant rooms and basements eat connectivity. Staff need to know whether the app stores the entry offline and syncs later, and that the time and location are still captured either way. Done when: an operative can log a reading in a dead spot and you can later prove when and where it was taken.
- Teach the responsible person to review, not just collect. A green dashboard is not oversight. The reviewer’s job is to find the missed flush in week three and the outlet that has read suspiciously identical for a month. Done when: the responsible person can surface every missed task and open exception for the past month in a couple of minutes.
- Build a short induction for relief and agency staff. People who missed the rollout still have to keep the trail intact. Done when: a new or temporary operative can work their first shift with their own login and a restricted task set, without anyone reaching for a shared account.
Where training quietly breaks
Even a good rollout drifts. A few failure modes are worth naming, because they rarely surface until an audit does.
The dashboard illusion is the big one. Averages and completion percentages flatter a system; a 98% completion rate can still hide the single low-use shower that has not been flushed since spring. Train reviewers to hunt exceptions, not admire totals.
Copy-forward habits are the second. When last week’s reading pre-fills this week’s field, tired people accept it. Identical figures down to the decimal, week after week, are a tell an auditor will spot before you do — and a pattern worth a spot-check.
Photos earn their place only if people know what a useful one looks like: the gauge with a legible number, the outlet being tested, not a blurry corner of a cupboard. The same goes for free-text notes — “flushed, fine” helps nobody in two years’ time. Many of these are the same habits that wreck paper systems, dressed in a new interface; Common mistakes in Legionella record keeping catalogues the recurring ones in Legionella record keeping.
Proving the training landed
You verify training the way you verify any control: with evidence, not optimism.
Pull a sample of records a fortnight after go-live and read them as a stranger would. Could someone who was not there reconstruct what happened from the entry alone? Then test the exception path — review a real out-of-range event, or stage one — and check the escalation actually fired and reached a named person. Finally, ask one operative the question an auditor eventually will: show me the last three flushes on this outlet, and what happened the time one was missed. If they can answer from the system in under a minute, the training worked. If they reach for memory, it didn’t.
How often you repeat any of this is set by your risk assessment and the rhythm of your monitoring, not by the software’s reminder defaults [2].
A note on limits
The tasks you train on, the thresholds you set in the app, and the routes an alert travels all come from your written scheme and a competent, site-specific risk assessment — the logbook records those decisions, it does not make them. Record-retention periods, monitoring frequency and the right remedial action depend on your system and the people it serves, so confirm them against current HSE guidance and your own assessment rather than copying another site’s settings [1].
Where to start
Pick one role and one asset this week. Sit with the operative who logs it, watch them record a normal reading and then a deliberately failed one, and read the two entries back cold. If the record alone would not tell an outsider what was checked, what the result was, and what was done about the bad one, your training gap is right there — in the judgement, not the menus. Fix that one record, then run the same test across the other roles.
FAQ
The software is meant to be intuitive — do we really need formal training?
Intuitive design helps people find the right button; it does not tell them what a reading should be or what to do when it fails. The duty holder is responsible for making sure whoever carries out the checks is competent to do them, and that competence is about judgement, not screen literacy [1].
How do we stop agency or relief staff breaking the audit trail?
Give them their own login, restrict them to the tasks they actually perform, and run a short induction before the first shift. Never hand out a shared account to save time — the moment two people log as one, the trail stops telling you who did what [1].
How often should we refresh the training?
Tie it to your risk assessment review, and bring it forward whenever something changes: new staff, a reconfigured system, revised control limits, or a run of record gaps that suggests people have stopped understanding what they record [2].
Sources
[1] HSE, “Legionnaires’ disease. The control of legionella bacteria in water systems - Approved Code of Practice and guidance (L8)”. https://www.hse.gov.uk/pubns/books/l8.htm [2] HSE, “Legionnaires’ disease: Technical guidance (HSG274)”. https://www.hse.gov.uk/pubns/books/hsg274.htm