Office Wi-Fi that survives Monday morning

You can cover an open-plan floor with very few access points. Getting a signal everywhere is not hard. What is hard is carrying what everybody is doing on that floor at the same time, from the same handful of radios, without anybody noticing each other.

Device density has quietly tripled

Count what one desk now generates: a laptop, a phone on Wi-Fi rather than mobile, sometimes a tablet, sometimes a headset with its own radio, and a growing number of building-services devices — sensors, desk-booking panels, digital signage, door controllers, printers, air-quality monitors — that all sit on the same spectrum. A design based on headcount will undercount the client population substantially, and a design based on floor area will not count it at all.

Video calls changed the load profile

The pre-2020 office workload was bursty and mostly downstream: fetch a page, fetch a file, go idle. A floor full of video calls is close to the opposite — sustained, symmetrical, latency-sensitive and jitter-intolerant, running continuously for an hour at a time. Wi-Fi handles bursty traffic gracefully by interleaving it; sustained bidirectional streams from many clients on one radio is a much less forgiving load, and it is the load that now defines the peak. Airtime, not bandwidth, is usually what runs out first.

Glass, plasterboard and the things that surprise people

Plasterboard partitions are almost transparent, which is why open-plan offices are easy to cover and why signal spreads much further than an architect's plan implies — often into a neighbour's floor plate. Glass is the interesting one: plain internal glazing is fairly transparent, but structural glazing, laminated partitions, and particularly the metallised or low-emissivity coatings on modern external glass, attenuate heavily. That coating is why an office can have superb coverage indoors and effectively none on the terrace, and why external access points cannot be relied on to serve the floor behind the glass. Solid cores, lift shafts, risers, plant rooms, wet areas and fire-rated partitions are the other consistent surprises.

Neighbouring tenants

In a multi-let building or a business park you do not have the spectrum to yourself. The floor above and the units either side are radiating into your space, and every one of them has almost certainly been configured at maximum power to fix a complaint. Turning your own power up in response is the classic escalation and it makes the shared environment worse for everyone, yourself included. The correct response is usually a denser design at lower power, with a channel plan drawn up in full knowledge of what the neighbours are doing.

Hot-desking and agile working

Flexible working has made occupancy far less predictable. The floor that is a third full on Monday is at capacity on Wednesday, and people cluster — around the good desks, the quiet corners, the areas near the coffee. A design that spreads capacity evenly across a floor plate on the assumption of even occupancy will be over-provisioned in some zones and short in exactly the ones people prefer.

Count devices, not people, and model the load

We establish the actual client population — user devices, mobiles, wireless peripherals, building services — and what those devices do at peak. Concurrent video is modelled explicitly because it dominates the airtime calculation. The output is a design whose capacity is deliberate rather than incidental. More on how that works on the Wi-Fi design page.

Smaller cells, lower power, more of them

The instinct in a dense office is fewer access points at higher power. The correct answer is nearly always the reverse: more access points, each covering less, at reduced transmit power, so more of the spectrum can be reused across the floor and fewer devices contend on each radio. This is also what makes 6 GHz worth having — considerably more usable spectrum, but with markedly less range, so a Wi-Fi 6E or Wi-Fi 7 design typically needs a denser access point layout than the 5 GHz network it replaces, not a sparser one.

Meeting rooms and collaboration clusters

Rooms get designed as their own capacity problem: the number of seats, whether the room is used for hybrid calls, whether there is a room system with its own wireless requirements, and whether the walls are glazed or solid. Clusters of small rooms along one wall are a distinct problem again — high density in a small footprint, often with the access point serving them from the corridor.

Ceiling voids, aesthetics and the compromise

Landlords, designers and heritage constraints all place real limits on where hardware can go, and those limits are legitimate. What is not legitimate is pretending they are free. We will show you what a concealed position costs in performance and offer the alternatives — low-profile enclosures, discreet colour-matched mounts, above-tile positions with the antenna below the line, or a different position altogether that achieves the same coverage without the compromise. The decision is yours; it should just be an informed one.

Segregation and policy

Corporate, guest, voice, building-services and IoT traffic separated at design stage, with SSID structure, VLANs, bandwidth policy and captive-portal requirements specified rather than improvised later. Every additional SSID has an airtime cost, so the structure is kept as lean as the requirement allows.

Fit-out sequencing

Where the office is being fitted out, a predictive survey from the architect's drawings and the partition schedule lets access point positions, containment and cabling be coordinated with the other trades instead of retrofitted afterwards. That means the containment goes in with everyone else's, the cabling is in before the ceilings close, and the validation survey happens once the partitions are up and the furniture is in — because a design modelled on an empty shell has to be confirmed against the finished space.