UniFi WiFi surveys, design and installation
UniFi is the platform we are called in to rescue more than any other, and almost never because the hardware is bad. It is because it was bought by access point count instead of designed.
What UniFi is genuinely good at
UniFi's advantage is stark and easy to state: the cost per access point is a fraction of the enterprise platforms, and there is no recurring licence. A controller you host yourself — on a Cloud Key, a Cloud Gateway, a virtual machine or a spare server — manages the estate at no ongoing cost. For a single site with a fixed budget, that difference can be the whole project.
The hardware is better than its price implies. The current access points support Wi-Fi 6 and 6E, the mounting hardware is decent, the controller is genuinely pleasant to use, and the network stack is coherent if you also use their switching and gateways. In an office of eighty people with ordinary laptops and phones, a properly designed UniFi network will do everything asked of it and cost a third of the alternative.
That is a real endorsement, and it comes with a real condition: properly designed.
Why UniFi networks fail more often
Almost every failed UniFi deployment we are called to has the same history. Somebody counted the rooms, bought that many access points, mounted them where the cabling was easy, and switched them on. It worked on day one with six people in the building, and it stopped working when the building filled up.
Bought by count, not by design
Because UniFi is cheap, it invites the assumption that more access points is a strategy. It is the opposite. Too many access points on overlapping channels at high power is the single most common fault we measure — every radio shouting over its neighbours, clients clinging to the wrong one, and throughput collapsing precisely when the building is busy. Adding another access point to a co-channel interference problem makes it measurably worse. This is what troubleshooting usually finds.
Default power settings
Access points ship loud. In a small building, leaving transmit power high means every access point hears every other one, the channel reuse plan collapses, and devices roam badly because a distant access point still sounds strong enough to cling to. The fix is a measured power and channel plan — which is a survey output.
Roaming-critical fleets
If your devices are barcode scanners, forklift terminals, AGVs or wireless VoIP handsets, roaming behaviour is the whole requirement, and this is where UniFi is most often the wrong choice. It supports the roaming standards; the question is how the specific firmware on your specific device fleet behaves against it at speed, and that is answered by measurement, not by a datasheet. More on roaming for scanners and AGVs.
High density
A lecture theatre, a function room or a packed office corner is a capacity problem. UniFi can be designed for density, but the margin for error narrows sharply and the design has to be right — directional coverage, tight power, careful SSID hygiene, and a realistic count of what will actually be in the room.
Controller as an afterthought
A self-hosted controller is a real piece of infrastructure. If it lives on a spare desktop under someone's desk with no backup, the day it dies is the day the configuration is gone. It is a small cost to do properly and a genuine outage to do badly.
How we survey for a UniFi design
The survey is the same wireless site survey we run for any platform — the difference is that for UniFi it more often changes the answer, because the design margin is narrower.
- Measured access point count and positions on a scaled plan. On the rescue jobs, this number is regularly lower than what is already installed
- A channel and power plan — the two settings that turn a bad UniFi network into a good one without buying anything
- Model and antenna selection — indoor, in-wall, high-density or outdoor, and where a directional antenna solves a problem that a stronger radio makes worse
- PoE and cabling schedule — Wi-Fi 6E access points have a real power budget, and it is a switch decision as much as an access point one. We do the cabling in house
- Controller placement and backup, stated in the design rather than assumed
- Guest network design where there is one — SSID count, isolation and portal behaviour all cost airtime. Guest Wi-Fi and captive portals
- A validation survey after installation, measured against the design. More on validation
Where the measurements say UniFi will not meet the requirement, we say so and show you the plot that proves it — rather than selling you the platform we would rather install.
Fixing a UniFi network you already have
Before quoting anyone for replacement hardware, it is worth knowing that a large share of UniFi complaints are configuration and placement, not capability. The usual sequence is:
- An on-site survey of the network as it stands, measuring signal, signal-to-noise, co-channel interference, channel utilisation and retry rates — plus a spectrum sweep for the interference that is not Wi-Fi at all
- A findings report that separates the three possible causes: too few access points, too many in the wrong places, or the right access points configured badly
- A remediation plan — frequently "turn the power down, fix the channel plan, move three of them and remove two", which costs a day rather than a project
- Only where the measurements genuinely demand it, additional or different hardware, with the evidence attached
We have no reason to tell you to replace a network that can be fixed. The survey is the same either way.
SEGwifi is an independent Wi-Fi surveyor, designer and installer. Product names are the trade marks of their respective owners and are used here only to describe the equipment we survey, design and install. We are not affiliated with, endorsed by, or acting as an agent for any of them.
Common questions
Is UniFi good enough for a business?
For a great many businesses, yes — and it saves a lot of money. The honest caveat is that UniFi has less margin for a poor design than the enterprise platforms, so the survey matters more, not less. Where it struggles is high density, roaming-critical device fleets such as scanners and AGVs, and estates that need multi-site management.
We have UniFi and it keeps dropping. Do we need to replace it?
Usually not, at least not all of it. The most common causes we measure are too many access points at too high a power on overlapping channels, and positions chosen for cabling convenience. Both are fixable without new hardware. A survey tells you which category you are in before you spend anything.
Can you install UniFi if we buy the hardware ourselves?
Yes. Many clients buy their own UniFi kit and bring us in for the survey, design, cabling and installation. We will tell you before you order whether the models you have chosen suit the building.
UniFi or Meraki?
Budget and estate shape decide it more than radio performance. Meraki costs considerably more, carries a mandatory licence and gives a stretched IT team far better multi-site visibility. UniFi costs far less with no recurring fee and is entirely capable in a well-designed single site. We will price both against the same survey if it is genuinely a close call.
Does UniFi need a survey if it is only a small office?
A small simple office with a handful of people is often fine on judgement alone, and we will say so rather than sell you a survey you do not need. The threshold is usually density, difficult construction, or anything that has to keep working while people move around — at which point measuring is cheaper than guessing twice.
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.