
Zone water meters: catching loss before it hurts
Most water leaves the pipe long before it shows up on a bill. Zone-level metering breaks a network into measurable districts so the first signs of loss are visible in days, not quarters.
Technology · 10 September 2026
Every metering rollout starts with the same infrastructure question: build your own radio network, or put a SIM in every meter and use the mobile operator's? Both LoRaWAN and NB-IoT were designed for exactly this job – battery devices sending small messages for a decade. The differences that matter are elsewhere: who controls coverage, how the costs are shaped, and what happens at the edges of the map.
LoRaWAN runs on unlicensed sub-GHz spectrum (868 MHz in Europe). You install the gateways – or use a public LoRaWAN network where one exists – and from then on the network is yours: no SIM cards, no per-device subscription, hardware capex and that’s it. A single well-placed gateway covers from a couple of kilometres in dense city fabric to over ten in open terrain, and meter batteries routinely last 10–15 years.
The defining property is control. If a cellar full of meters has weak signal, you fix it yourself by adding or moving a gateway – today, not when an operator’s rollout plan gets around to it. The economics follow density: once a few hundred devices share an area, the cost per meter of owning the network drops below almost any alternative.
NB-IoT is a cellular technology on licensed spectrum. Every meter gets a SIM or eSIM, and the mobile operator’s network does the carrying – there is nothing to build. Its deep-indoor propagation is excellent, which is why it shines exactly where metering is hardest: basements, pits and chambers. Battery life is long for stationary devices, and a rollout can start the day the meters arrive.
The trade-offs mirror the strengths. Costs are a per-SIM subscription forever, which is unbeatable for a handful of scattered meters and adds up across tens of thousands. And coverage is the operator’s decision, not yours: where a pit has no signal, you wait or work around it – you cannot hang a gateway. The history of 2G and 3G sunsets is also a reminder that an operator’s technology roadmap is part of what you are buying.
For metering, most spec-sheet differences are noise – both technologies move a few small messages a day with ease. Four differences decide real projects:
Real networks are not uniform, so the right answer is rarely one technology. The pattern we see across rollouts is a deliberate split: LoRaWAN for the dense core where the density economics work, NB-IoT for the scattered edge cases, and walk-by or drive-by collection for the meters no network reaches economically. Asker municipality runs more than 20,000 meters as a LoRaWAN and wM-Bus hybrid on exactly this logic – technology chosen per area, not per organisation.
The precondition for deciding per area is a platform that does not care. In Divako, LoRaWAN, NB-IoT, wM-Bus and drive-by readings land in the same device library, the same data model and the same billing export – so the network choice stays an engineering decision you can revisit, meter fleet by meter fleet, instead of a commitment made once for everything. The full side-by-side of all five connectivity technologies is on our connectivity comparison page.
If the last answer is yes, the choice stops being scary – it becomes tuning.

Most water leaves the pipe long before it shows up on a bill. Zone-level metering breaks a network into measurable districts so the first signs of loss are visible in days, not quarters.

A worked ILI calculation: build the water balance, find the four UARL inputs and check the result with a practical network example.

Berg Insight lists Divako among software solution providers in its 2026 report on smart water metering in Europe and North America.
Your rollout
Tell us your device count and geography – where the meters cluster and where they scatter. We'll sketch which technology fits which area, in about 30 minutes.