Before a single sensor goes into a meter pit, a warehouse ceiling, or a city cabinet, there is one question a deployment team needs answered: will the signal actually hold at this exact spot? A network operator's published coverage map won't tell you that. LoRaWAN coverage testing done on site, with the gateway in its real mounting position, is the only reliable way to know before hundreds of devices are already installed and the truck has left.
Operator coverage maps are built from propagation models, not from the specific building, basement, or industrial site you're deploying into. They rarely account for the materials between a sensor and the nearest gateway — concrete, metal shelving, low-e glass, underground vaults — or for the exact height and orientation the gateway antenna will end up in once it's mounted. A location that looks solidly covered on a map can still fail once real walls, real inventory, and a real ceiling-mounted gateway are in the picture.
A proper site survey captures three things at every candidate location: RSSI (received signal strength, generally healthy above roughly -100 dBm), SNR (signal-to-noise ratio, ideally above -10 dB), and how many gateways actually hear the device. That last number matters more than most teams expect — a location reached by two or three gateways survives a single gateway outage or a temporary obstruction; a location reached by exactly one does not. Packet loss under real conditions, not just a single successful ping, is what separates a location that looks fine from one that will actually hold up in production.
Start by testing with the gateway in the position it will actually occupy — a unit sitting on a desk during the survey and later ceiling-mounted at 4 meters will report two different worlds. Walk or drive a grid across the deployment area, testing every 150–300 meters in dense urban or indoor environments and every 500 meters to a kilometer in open, rural terrain. Log RSSI, SNR, and gateway count at each point rather than relying on a single spot check near the entrance.
Test under conditions that resemble the finished site, not the empty one. A warehouse surveyed before pallets arrive, or a building surveyed before tenants move in, will read better than it performs once the space fills up. If the environment is prone to interference — a busy industrial floor, a dense residential block — repeat a handful of readings at a different time of day before locking in gateway placement.
Numbers on their own don't tell an installer what to do. The teams that get this right convert RSSI, SNR, and gateway redundancy into a single go/no-go signal for each location, rather than asking every field technician to interpret a spreadsheet of dBm values against a link budget. That's the gap between a coverage test that produces data and one that produces a decision — and it's the difference that determines whether a rooftop or basement gets a green light on the spot or a second visit next week.
Field testers span a wide range, from dedicated RF instruments built for specialists to phone-based tools designed for mixed teams of installers and engineers. Loptimizer, the LoRaWAN field tester from Pilot Things, is built for the second group: a Bluetooth Pocket Tester paired with a mobile app that converts device RSSI, gateway SNR margin, and gateway redundancy into a single 0–100 location score, GPS-tagged on a map and organized by project and floor.
If you're comparing testers before you buy, Loptimizer vs. the Competition on the Loptimizer blog walks through a feature-by-feature comparison against the Adeunis FTD, Glamos Walker, Watteco Netw'O, and Milesight FT101 — useful reading before you commit to hardware for a full deployment team.
A half-day site survey is cheap compared to the alternative: a second truck roll to relocate a gateway, or a sensor that drains its battery in eighteen months because it was installed at the edge of coverage and kept retransmitting. Testing coverage on site, with the real gateway position and real building conditions, is what turns a deployment plan into devices that stay online for years instead of months.
