On-site Wi-Fi surveys
Measured in the building, not modelled from a drawing. Walls as built, racking as loaded, and the neighbours' networks included — because all three change the answer.
Why measuring beats modelling
Buildings are not built exactly as drawn. Partitions move, materials get substituted, services get routed through the ceiling void the access points were meant to occupy, and somebody puts a steel mesh in a wall that the drawing showed as plasterboard. None of that appears in a model, and all of it changes the radio environment.
Then there is everything a drawing could never contain: the racking, once loaded with stock; the neighbouring business three floors down running eight access points at full power on your channels; the machinery that throws non-Wi-Fi interference across half the spectrum; the mezzanine somebody added last year.
An on-site survey measures all of it, in place, as it actually is. It's the difference between a design that should work and a design you can hold someone to.
Further reading: Predictive vs on-site surveys, honestly compared — what measuring gives you that modelling cannot.
What happens on the day
Before we arrive
We agree a scaled plan of the site, the areas in scope, access arrangements, any inductions or permits, and the device types and applications the network has to support. If there's an existing network we'll want its configuration — or we'll survey it as we find it.
On site
The surveyor walks the site with calibrated survey equipment and the scaled plan, so every reading is positioned accurately on the drawing. We capture signal strength, signal-to-noise, channel utilisation, interference (including non-Wi-Fi sources through spectrum analysis), data and retry rates, and roaming behaviour along the routes people and devices actually take.
The core of it is the AP-on-a-stick method. A survey access point goes up on a tripod at the first position the plan suggests, and the surveyor walks the surrounding area recording real measurements. That position is then fixed in the model and the process repeats for the next one, until every candidate position has been measured rather than predicted. The results combine into one picture of how the finished network would behave with all the access points working together.
It is slower than modelling and it is the reason the design survives contact with the building. It settles the arguments a model cannot: high-bay warehouses, atria, curved and listed buildings, and anywhere the predicted answer looks marginal.
Access, disruption and safety
Surveying is passive — we're listening, not reconfiguring, so nothing goes down. We do need physical access, including at height and into plant areas, and we'll need to know about restricted or hazardous zones in advance. Occupied offices, live warehouses, schools in term time and out-of-hours working are all routine. RAMS are provided as standard, and our engineers carry the site paperwork you'd expect through The Specialist Electrical Group's accreditations.
Where an on-site survey earns its money
- Warehouses and distribution centres. Racking attenuation changes with what's on the racking. A survey of an empty warehouse and a survey of a full one are different surveys.
- Older and listed buildings. Solid walls, unexpected materials, and restrictions on where anything can be fixed.
- Dense urban sites. Where the biggest variable isn't your building, it's everyone else's network.
- Sites with machinery or plant. Non-Wi-Fi interference that no model would predict and no phone app would spot.
- Anywhere roaming matters. Scanners, voice handsets and AGVs need coverage overlap designed deliberately and measured in place along the actual route.
- Before a hardware refresh. Measuring first regularly shows that the existing kit was never the problem.
Common questions
How long will you be on site?
A single-floor office is usually a day. A large distribution centre or multi-building estate is several, and outdoor areas add time. Racking density and access arrangements drive the number more than floor area does — you'll get a timescale before you commit.
Does anything need to be switched off?
No. The survey is passive measurement. If we run an AP-on-a-stick exercise we bring our own access point and power; it doesn't touch your network.
Can you survey out of hours?
Yes, and sometimes we should. A retail floor or a school is easier to survey when it's empty — but a capacity problem can only be measured when the people are there. We'll usually want both: an empty-site pass for the physics, and a loaded pass for the behaviour.
Do we need to be there?
Someone who knows the building needs to be available at the start — where the plant rooms are, which doors are locked, what's behind the wall nobody talks about. After that the surveyor can generally work unaccompanied, subject to your site rules.
What if the site changes after the survey?
Tell us and we'll say whether it matters. New racking, a mezzanine, a partition layout change or a different device fleet can all shift the design. Minor changes are usually a design revision; major ones may need a re-survey of the affected area, not the whole site.
Let's find out what your Wi-Fi is actually doing.
Book a survey and you get measured data, a design you can build to, and a number you can budget against — not an opinion.