A remote monitoring system has one job: turn continuous temperature data into a decision you can defend. Not a wall of dashboards, not a heat map nobody opens. If a sensor reading never triggers an action or proves a control held, you have bought scenery.
So before you compare brands, fix what the purchase must achieve. You want fewer manual rounds, faster reaction when a calorifier drifts, and an evidence trail that survives an audit without a fortnight of spreadsheet archaeology. Everything below is graded against those three outcomes.
This is a commercial decision with a compliance edge — exactly where buyers overspend on features and underspend on the parts that decide whether the thing gets used.
What the system has to deliver
Three things, in order. It has to measure the right points accurately. It has to tell the right person when something is wrong, in time to matter. And it has to leave a record you can hand to an assessor.
A sensor count is not a benefit. A reduction in unnoticed temperature excursions is.
The buying framework: four criteria and the questions that test them
Score each vendor against these four. Weight them for your site — a single small building cares less about integration than an estate of forty does — but do not let a slick interface paper over a weak sensor.
1. Sensor accuracy and calibration
Everything downstream depends on the number being right. A clamp-on pipe sensor reads differently from a sensor in the flow; both can be useful, but you need to know which you are getting and what it is actually measuring. HSG274 expects monitoring at sentinel and representative outlets chosen by your risk assessment, so the kit has to be placeable where your assessment says it matters, not only where wiring is easy [1].
Ask the vendor: what is the stated accuracy in degrees, and over what range? How is each sensor calibrated, how often, and who pays for recalibration? Can you show me the calibration certificate trail an auditor would want? What happens to a reading when a sensor fails — does it alert, or does it silently flatline? In my view, a system that cannot prove a flatlined sensor failed is worse than no sensor, because it manufactures false confidence.
2. Alerting logic
An alert that fires too often gets muted, and a muted alert is no alert. The value is not in detecting an excursion; it is in routing it to someone who will act before the next person draws water.
Ask: can thresholds differ per outlet, so a sentinel cold outlet at 20°C and a stored-hot calorifier near 60°C each get sensible limits rather than one blunt rule? HSE frames cold water as kept below 20°C and hot stored around 60°C with roughly 50°C at the outlet, but your risk assessment sets the operating figures, so the system must let you encode that nuance [2]. Can you set escalation — if no one acknowledges in an hour, it goes up a level? Can it suppress the predictable nuisance alert (a flushing event, a known low-use outlet) without you switching off the genuine ones? What channels: email, SMS, app, a logbook task?
3. Integration with your records and estate
Continuous data that lives in a separate portal becomes a second system to reconcile, and reconciliation is where compliance quietly rots. The monitoring feed should land in the same place as your manual checks, asset register and remedial tasks, so one record tells the whole story.
Ask: does it export or push into a digital logbook, or is it an island? Is there an API, and is it documented and included, or a paid extra? If you run a BMS, does it speak a protocol the BMS understands? Can a temperature breach automatically raise a remedial task with an owner and a due date, or does someone have to retype it? The pragmatic call: prefer a system that writes to your record of truth over one with a prettier dashboard of its own.
4. Support and the service model
Sensors drift, gateways drop off the network, batteries die. What you are really buying for years two through five is the support behind the hardware.
Ask: what is the response time when a gateway goes offline, and is connectivity monitored at their end or only noticed when you log in? Who replaces a failed sensor, and on what timescale? What is the total cost over five years — hardware, connectivity, software licence, recalibration, replacements — not the headline install figure? When the contract ends, do you keep your historical data, and in what format?
A vendor question table to take into the demo
| Criterion | The question that exposes weakness | A weak answer sounds like |
|---|---|---|
| Sensor accuracy | ”Show me a calibration certificate and the stated accuracy band." | "It’s very accurate” with no number |
| Sensor failure | ”What happens on screen when a probe fails?" | "That rarely happens” |
| Alerting | ”Set me a per-outlet threshold and an escalation rule, now." | "All alerts use one default limit” |
| Integration | ”Push a breach into my logbook as a task with an owner." | "You can read it in our portal” |
| Data ownership | ”At contract end, what do I keep and how?” | Vague, or an export fee |
| Five-year cost | ”Itemise every recurring charge over five years.” | Only the install price |
If a vendor cannot do these live, that tells you more than any brochure.
Trade-offs worth naming
More sensors mean more coverage and more maintenance. Wireless installs fast but depends on signal in plant rooms with thick walls; wired is harder to retrofit. Tighter alert thresholds catch more but cry wolf more. There is no free setting here, only a balance your risk assessment should inform.
Red flags
A demo that only shows the happy path and never a failed sensor. Accuracy quoted as a percentage with no degree figure. “Compliance guaranteed” language — no product delivers that; your scheme does. Data you cannot export. A locked ecosystem where every integration is a quote. And a price that looks cheap because connectivity, licence and recalibration are billed separately later.
When not to buy
If a single small building with a handful of outlets is well served by a competent person and a monthly round, continuous monitoring may add cost without changing a decision. The case strengthens with scale, hard-to-reach plant, out-of-hours risk, and where manual rounds keep slipping. Buying kit to fix a process problem — missed checks, unclear ownership — usually just digitises the gap.
This is general guidance, not a design specification for your premises. What you monitor, where, and at what thresholds must come from a competent, site-specific Legionella risk assessment under ACoP L8 and HSG274 — a monitoring system supports that scheme; it does not replace the duty holder’s judgement [1][3].
Your next step today: list your sentinel and representative outlets from the current risk assessment, mark which ones you cannot reach easily or check reliably, and take that exact list into your first vendor demo. If your records still live on paper or a spreadsheet, decide where the monitoring feed would land first — moving to a digital logbook before you add sensors means the data has a home the day it arrives.
FAQ
Does remote monitoring satisfy my legal monitoring duties on its own?
No. It is a tool that supports your control scheme. The duty holder still owns the risk assessment, the wider control measures and the decisions; HSE guidance treats monitoring as part of managing the system, not a substitute for it [3].
How accurate does a temperature sensor need to be?
Accurate enough that a reading near your action limit is trustworthy. Ask for the stated band in degrees and the calibration trail rather than a vague claim. The exact figures that matter come from the limits your risk assessment sets against HSE’s hot and cold expectations [2].
Should the monitoring system replace my manual checks?
Usually it reduces them rather than removes them. Some checks need a person — a flush, a descale, a look at a TMV — that no sensor performs. Decide which rounds the data genuinely makes redundant and keep the rest.
Sources
[1] HSE, “Legionnaires’ disease: Technical guidance (HSG274)”. https://www.hse.gov.uk/pubns/books/hsg274.htm [2] HSE, “Hot and cold water systems”. https://www.hse.gov.uk/legionnaires/hot-and-cold.htm [3] HSE, “Testing and monitoring your water system for legionella”. https://www.hse.gov.uk/legionnaires/testing-monitoring-water-system.htm