Wi-Fi Site Survey Guide: Measure Before You Drill

An AP in the wrong place doesn't just mean weak signal. It means new holes in the ceiling, new cable pulls and patching the old holes. In a hotel with guests in residence or an office already in use, that kind of fix is expensive and disruptive.

Usually the cause is guesswork. Concrete walls, low-E glass, steel racking, tiled bathrooms and lift shafts absorb or reflect Wi-Fi in ways you can't see on a floor plan. A site survey tells you how many APs you need, where to mount them and how to configure them, with measurements to back it up.

Four kinds, and which to use

In practice, a good project combines two: predictive during design and validation at handover. Add AP-on-a-stick for buildings with unusual construction or where the stakes are high.

The numbers we design to

Why does voice need two APs? Because people walk while they talk. The phone needs a second AP to roam to before the first one fades, or the call drops mid-sentence.

Beyond RSSI and SNR, you also want minimal co-channel overlap (APs on the same channel that can hear each other have to share airtime) and a realistic number of clients per AP.

Design for the weakest device

Targets are set by the worst radio on the network, not the best laptop. Handheld barcode scanners, Wi-Fi phones and older tablets often hear the AP well but transmit weakly, so the AP cannot hear them back. Tell the surveyor which devices matter and, if possible, lend one for the survey day.

Capacity: a worked example

A conference hall seats 200 people. Assume most bring one active device for slides, messaging and the occasional video, at around 2 Mbps each while active. That is roughly 400 Mbps of demand across the room. For this kind of use we plan around 30–40 active clients per 5 GHz radio, so the hall needs five to seven APs, on 20 MHz channels so neighbouring APs don't share the same channel, with transmit power turned down so each AP serves its own part of the room. Coverage alone would have said two APs were enough. That gap is why halls that look fine on a heatmap fail on the first full day.

6 GHz changes the survey

Wi-Fi 6E and Wi-Fi 7 add the 6 GHz band, which gives far more clean channels but loses more energy through walls than 5 GHz. A design that works on 5 GHz can leave 6 GHz patchy in the same rooms. If you are buying 6 GHz-capable APs, survey with a 6 GHz-capable adapter and check both bands in the heatmaps. Also confirm the AP's country code is set for where it is installed; it decides which channels and power levels the AP may use, and units bought for another market are not always set correctly.

What survey day looks like

Before the visit we ask for scaled floor plans, the priority areas (meeting rooms, lobby, restaurant), concurrent user counts and the applications that matter.

On site the engineer walks the building and checks real materials against the drawings. They rarely match: a wall drawn as gypsum turns out to be brick, a ceiling assumed to be plasterboard turns out to be metal panels. We check ceiling height, viable mounting points, and cable paths that stay within 90 m of the switch.

Interference scan. Neighbouring Wi-Fi (on some floors in central Vientiane you'll see 30–40 SSIDs), microwaves, wireless cameras and other 2.4 GHz devices.

Design. AP positions, channels, channel width (20 or 40 MHz in dense areas) and transmit power.

After installation we run a validation survey and hand over the heatmaps.

Your preparation checklist

• Floor plans to scale, with the scale or one known dimension marked
• Ceiling access arranged, and an escort for restricted areas
• A list of the client devices that matter, including the oldest
• Peak user counts per area and the times they occur
• Permission for the surveyor to mount a temporary AP on a stand or ceiling
• Any planned changes: new glass partitions, racking, extra floors

Where surveys go wrong

Designing for coverage and forgetting capacity. A hall with full signal bars and 200 people in it still won't work.

Predictive only, with no site visit. Real materials and drawings often disagree.

Measuring with a phone app. Phone antennas vary widely between models, and most apps don't measure SNR or co-channel overlap at all.

Surveying an empty building. People are mostly water, and water absorbs 2.4 and 5 GHz well. A full meeting room measures noticeably weaker than an empty one.

And skipping validation. Without it there's no proof the targets were met, and nothing to compare against when problems are reported.

One configuration mistake shows up in almost every survey of an existing network: 2.4 GHz switched on at full power on every AP. There are only three non-overlapping 2.4 GHz channels (1, 6 and 11), so in a building with twenty APs most of them interfere with each other. Turn 2.4 GHz off on a share of the APs, lower the power on the rest, and keep that band for the older devices that need it.

The report you should receive

Keep it. When desks move, headcount grows or someone says the Wi-Fi is slow, you have a baseline to compare against and can extend the design without starting over.

By site type, in Laos

New buildings in Vientiane: run a predictive survey during design so the cable to each AP location is in before the ceiling closes. Adding drops afterwards costs far more.

Hotels and resorts: AP-on-a-stick. Bathroom walls and glass make a big difference, and no two buildings behave the same.

Warehouses and mines: survey when the racks are full or the machinery is running. An empty warehouse and a full one are completely different RF environments.

High-bay warehouses with ceilings of 10–12 m usually need APs mounted lower, on the racking or on drop poles, often with directional antennas pointed down the aisles. Omnidirectional APs at roof height light up the roof, not the scanners at floor level.

Outdoor areas, rooftops and site camps need outdoor-rated APs, and every copper PoE run that leaves the building needs surge protection. The survey should also note where the rain season changes things: wet foliage between buildings absorbs signal on point-to-point and outdoor links.

In the contract, make a validation report a condition of acceptance.

When to survey again

A survey is a snapshot. Plan a new validation walk when:

• Walls, glass partitions or shelving are added or moved
• Headcount in an area grows noticeably, or a room changes use
• You replace APs with a new Wi-Fi generation or add 6 GHz
• Users report the same problem in the same place more than once
• A neighbour installs a large new network next to yours

Before you order access points

Send us your floor plans (PDF, or a photo where the scale is readable) and tell us which areas matter most. We'll do a first predictive pass and tell you whether your site needs on-site measurement. For a rough number in the meantime, try our Wi-Fi AP estimator.