Wi-Fi specialists — we survey it, design it, install it, and prove it works.
Home / Wi-Fi design

Wi-Fi design

Design is where a survey becomes a network. Access point placement, channel and power planning, capacity modelling, roaming and segregation — all worked out from measured data rather than a rule of thumb.

What Wi-Fi design actually involves

There is a widespread and expensive belief that designing Wi-Fi means dividing the floor area by a number and buying that many access points. It is how a great many networks get specified, and it is why a great many of them fail under load.

Real design is a set of trade-offs made deliberately. Every access point you add gives you capacity and takes away spectrum from its neighbours. Every decibel of transmit power extends coverage and increases the area over which devices contend with each other. Coverage designed for a laptop will not hold a handheld scanner's session together. Getting those balances right in a specific building, for a specific set of devices, is the work.

Adding access points without a channel plan makes things worse Same channel, full power ch 36 ch 36 ch 36 Three radios all waiting for the same airtime. More hardware, less usable capacity. Planned channels, managed power ch 36 ch 44 ch 52 Overlap kept deliberate — enough for roaming, not enough to contend.
The most common self-inflicted Wi-Fi fault. Access points on overlapping channels at full power do not add capacity — they take turns on the same airtime and add noise to each other. The fix is usually a channel and transmit-power plan rather than more hardware.

The problems good design solves

  • Capacity that collapses at peak. Designing for coverage alone produces a network that looks perfect and fails at nine in the morning.
  • Access points fighting each other. Too many radios, too much power, overlapping channels. The most common self-inflicted fault we find.
  • Sessions dropping on the move. Roaming has to be designed — deliberate overlap, correct thresholds, the right band steering — or scanners and voice handsets will cling to a distant access point until they fail.
  • Guest traffic degrading the business. Segregation, bandwidth policy and airtime fairness, decided at design stage rather than patched afterwards.
  • Designs that can't actually be installed. An access point position is only valid if something can be fixed there, power and data can reach it, and it can be reached again for maintenance.

What we design

Coverage and capacity

Access point positions, mounting heights and orientation, worked to the signal thresholds your specific device fleet needs — with separate targets where you run mixed fleets. Capacity is modelled against real device counts and what those devices actually do, because forty people on video calls and four hundred barcode scans a minute are very different loads.

Further reading: Wi-Fi 6E and 7 in practice — what the newer standards change in a real building, and what they do not.

A scaled office floorplan marked with four proposed access point positions
The design is a drawing, not a shopping list. Every access point has a position on a scaled plan of your building before anyone quotes for a single mount or a metre of cable.

Channel and transmit-power plan

Which channel and what power for every radio, including how the design coexists with the networks in neighbouring units — and, in dense buildings, whether to compete for spectrum or deliberately cede some. This is handed over as a commissioning sheet, so it survives contact with whoever configures the kit.

Roaming

Designed overlap along the routes devices actually travel: aisles, corridors, stairwells, loading bays, yards. Roaming thresholds, band steering and minimum data rates set for the least tolerant device on the site, not the most.

Segregation and policy

Corporate, guest, voice, IoT and building-services traffic separated at design stage, with VLANs, bandwidth policy and captive-portal requirements specified rather than improvised. In care and healthcare settings this also means designing around nurse-call, telecare and medical equipment that shares the environment.

The physical layer

Cable routes, containment, cabinet positions, switch port counts, PoE budget and power. This is the part most Wi-Fi designs leave to somebody else, and it is the part that most often derails a programme. We design it because we install it.

Vendor independence, and why it matters

We work regularly with Cisco Meraki, Ruckus CommScope, Cambium Networks, Ubiquiti UniFi and more. We are not obliged to specify any of them.

That matters because the honest answer to "which platform should we use" is usually "several would work, and here is what separates them for your situation" — licensing model, management overhead, outdoor range, how well it handles high density, whether your team can administer it, and what happens at renewal. A design that starts from a preferred vendor and works backwards to a justification is a procurement exercise wearing an engineering costume.

Where you already own a platform, the design will work with it wherever it can genuinely do the job — and where it can't, you'll get the measurements that prove it rather than an assertion.

What you get

  • Access point schedule: position, height, orientation, fixing detail, model and part number
  • Coverage and capacity plots showing the designed outcome, at stated thresholds
  • Channel and transmit-power commissioning sheet
  • Roaming and segregation specification, including SSID and VLAN structure
  • Cabling, containment, cabinet and PoE schedule
  • Costed kit list with part numbers, so the work can be tendered on equal terms
  • A plain-English summary for the people approving it

The design is yours. If you want somebody else to install it, it is written so they can.

Roaming has to be designed, not assumed Insufficient overlap — the session drops still clinging to AP-01… …drops here walking route Designed overlap — clean handover overlap zone stays connected
Roaming is decided by the device, not the network. A handheld scanner or voice handset will hold onto the access point it is associated with until the signal becomes unusable — so if the cells do not overlap by design, the session fails before the device thinks to move. This is why adding access points without correcting overlap and roaming thresholds so often changes nothing.
FAQ

Common questions

Can you design without doing a survey first?

We won't, and you shouldn't want us to. A design without measured or modelled data behind it is a guess with a document wrapped round it. Where a building doesn't exist yet, a predictive survey provides the data.

Will you design around our existing hardware?

Yes, wherever the hardware can genuinely do the job. A significant share of the underperforming networks we see have entirely capable equipment that was never commissioned properly — in those cases the fix is design and configuration, not capital spend.

How many access points will we need?

Nobody can answer that responsibly without data. Anyone who quotes an access point count from a floor area over the phone is guessing, and the guess is usually low — which is how a quote wins and a project disappoints.

Do you design for Wi-Fi 6E and Wi-Fi 7?

Yes. Be aware that 6 GHz trades range for spectrum: it carries far more capacity but is much less tolerant of walls and distance, so a 6E or Wi-Fi 7 design typically needs more access points than the 5 GHz design it replaces, not fewer. That's precisely the sort of thing worth establishing before committing to a refresh.

Can you design for outdoor areas as well?

Yes — yards, car parks, campuses, quaysides and open ground. Outdoor design brings in weatherproofing, mounting and lightning protection, power and containment to remote positions, and different propagation behaviour. It's specified and priced separately.

Who owns the design?

You do. It's written to be tendered, and we're content to compete for the installation on price and capability rather than on holding your specification hostage.

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.