Wi-Fi troubleshooting and optimisation
Something is wrong and nobody can say what. We measure it, isolate the cause and prove it — before anybody spends money. Often the fix is engineering time, not new equipment.
Symptom, cause, and what it usually turns out to be
Wi-Fi complaints arrive as symptoms. The useful work is turning a symptom into a measured cause, because the obvious remedy is frequently the wrong one — and "add more access points" is the most expensive wrong answer in this industry.
| What people report | What it usually turns out to be |
|---|---|
| "It drops in the far aisle" | Roaming, not coverage. The device is clinging to a distant access point because the overlap and thresholds were never designed. Adding access points without fixing the thresholds changes nothing. |
| "Full bars, nothing loads" | Capacity or noise. Either the airtime is saturated, or the signal-to-noise ratio is poor because of neighbouring networks or non-Wi-Fi interference. Signal strength is fine and irrelevant. |
| "It's fine in the morning, terrible by ten" | Load. The design counted floor area rather than devices and applications, so it works empty and fails full. |
| "It got worse after we added access points" | Co-channel contention. More radios on overlapping channels at full power means more devices competing for the same airtime. This is a very common self-inflicted fault. |
| "Only the scanners have a problem" | Design target. Handheld scanners and voice handsets need higher signal levels and designed roaming overlap than laptops do. A network built to a laptop threshold will fail them and nothing else. |
| "It's fine until the shutter opens" | Environmental. Loading bays, shutters, moving stock and vehicles change the RF environment through the day. Only measurable in place, under real conditions. |
| "Guest Wi-Fi is killing the business network" | Segregation and airtime policy. Separate SSIDs on the same radios still share airtime unless the policy says otherwise. |
| "Nobody can tell us what's wrong" | Nobody has measured it. Two suppliers, two opinions, no data. |
How a troubleshooting survey runs
A troubleshooting survey is diagnostic rather than design-led, so it's structured differently to a standard site survey.
- Reproduce the problem. When does it happen, where, to which devices, doing what? Precision here saves days — "the Wi-Fi is bad" and "voice handsets drop calls between the packing benches and the loading bay after 3pm" lead to completely different investigations.
- Measure under real conditions. Coverage, signal-to-noise, channel utilisation, retry and data rates, roaming behaviour along the affected routes — during the period the fault occurs, not when the site is empty.
- Analyse the spectrum. Non-Wi-Fi interference doesn't appear as a network and won't show up in any tool that only sees Wi-Fi. Plant, machinery, and equipment nobody thought to mention are all candidates.
- Check the configuration. Transmit power, channel allocation, roaming thresholds, minimum data rates, band steering, SSID count and airtime policy. A great deal of poor performance lives here rather than in the physics.
- Look past the radio. Sometimes the Wi-Fi is fine and the problem is an uplink, a switch, a DHCP scope or a firewall. Where that's the case we say so plainly rather than selling you access points.
- Report with a remedy. Cause, evidence, and specific actions ranked by impact — including which ones cost nothing.
Optimisation without new hardware
A meaningful proportion of the sites we survey have capable hardware that has never been commissioned properly. Every radio on automatic and drifting to maximum power. Channels overlapping. Roaming thresholds untouched since the box was opened. Legacy data rates still enabled, dragging the whole cell down. Six SSIDs broadcasting when two would do.
Where that's the situation, the fix is engineering time. We re-engineer the configuration against measured data, then validate that it worked. It is a much smaller number than a refresh, and we would rather tell you that than sell you equipment you don't need — partly because it's honest, and partly because the client who was told the truth is the one who rings back.
Where the hardware genuinely can't do the job — too few radios, wrong generation for the device fleet, no capacity headroom — you get the measurements that prove it, which is a far better basis for a capital request than an opinion.
Common questions
How fast can you get to site?
Tell us how urgent it is when you call. Operationally critical faults — a warehouse that can't pick, a venue with an event tomorrow — get prioritised. We'd rather be straight with you about availability than promise a response we can't hold.
Can you help if you didn't install the network?
Yes, that's most of this work. We survey and report on what's there regardless of who installed it or what platform it runs on.
What if the problem isn't the Wi-Fi?
We tell you. Uplink saturation, switching, DHCP, DNS and firewall problems all present as 'the Wi-Fi is bad'. Where the radio is healthy, the report says so and points at what isn't — that's a useful outcome even though it doesn't sell us anything.
Will we have to replace all our access points?
Often not. A good share of these jobs end in a configuration change rather than a refresh. Where hardware does need replacing you'll get the measurements that justify it, not an assertion.
Do you charge for a troubleshooting survey?
Yes — it's skilled diagnostic work and it takes survey-grade equipment and time. The scoping conversation beforehand is free, and it sometimes resolves the problem on its own.
What if the fault is intermittent?
Those are the interesting ones. Intermittent faults usually correlate with something: a shift pattern, a shutter, a machine, a delivery, a time of day. We'll want to be on site when it happens, and in some cases we'll leave monitoring in place to catch it.
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.