Finding Dead Zones with Data (Not Just Complaints)
A dead zone is any spot where your wireless coverage is too weak to use. It announces itself in the least useful format known to IT: vibes. "The Wi-Fi is bad over there." "It never works in the back corner." "I just don't take my laptop to that end of the building anymore." No error messages, no device names, no timestamps. Just a slow leak of user trust, in an area you may not visit from one month to the next.
Here is the reframe that makes the problem tractable: coverage is a measurable, mappable physical fact. Radio signal at any point in your building has a number, the number can be recorded, and the recordings can be drawn on a floor plan. Better still, you already own two free data sources — the complaints themselves, and the per-client signal readings your access points collect all day. The third source costs only a walk around the building with a laptop.
This guide turns those three sources into a coverage map. It covers harvesting the complaint pattern, mining what the infrastructure already knows, walking a simple survey, and drawing the results. It covers fixing the holes you find and — the step everyone skips — proving the fix with the same walk. By the end, "the Wi-Fi is bad over there" becomes a colored spot on a map with a plan attached.
In this guide
- The complaint map: your first survey
- Free survey data your APs already have
- The walk survey, step by step
- Putting results on the floor plan
- Reading the map: holes, shadows, and illusions
- Fixing the hole
- Validating with the same walk
- The living coverage map
- Where to go next
The complaint map: your first survey
Before touching a tool, mine the data you have been throwing away. Every Wi-Fi complaint has a location attached — usually implied, occasionally stated — and locations are exactly what a coverage investigation needs. Go through recent tickets and hallway grumbles and tally them by place, not by person or date. Add one question to your helpdesk intake — where were you standing when it happened? — and the tally builds itself from now on.
A month of located complaints usually looks something like this:
| Area | Complaints | Typical wording | Smells like |
|---|---|---|---|
| NE corner offices | 7 | "No signal," "unusable," "I gave up" | Coverage hole — survey first |
| Conference room B | 5 | "Dies when I walk in mid-call" | Roaming, not coverage |
| Cafeteria | 4 | "Slow at lunch" | Congestion, not coverage |
| Stairwell landings | 2 | "Drops between floors" | Expected physics; probably a shrug |
Two lessons hide in that table. First, complaints cluster — a genuine hole generates repeat reports from one area, and the cluster tells you where to walk first. Second, not every cluster is a dead zone. "Dies when I walk in" is a roaming story and "slow at lunch" is an airtime story. Each has its own guide and its own fix. The complaint map's job is triage — separating where it is weak from where it is busy from where handoffs fail — so you spend your walking time where walking helps.
One bias to correct for: complaints require complainers. A storage floor with two occupants can have the worst coverage in the building and a spotless ticket history. Guest-facing areas underreport because visitors suffer silently and leave. Silence in the complaint map is no data, not good news — which is exactly why the map is the first survey and not the last word.
Free survey data your APs already have
Your second free source runs around the clock: every access point continuously measures the signal strength it hears from each associated client. Most APs and controllers will show you that list on demand. And it is a survey conducted by your users' own devices, from wherever they actually sit, updated all day, at a cost of zero.
ap-fl1-northeast associated clients: 9 client ...:4a2f signal -58 rate high client ...:9c11 signal -62 rate high client ...:77d0 signal -81 rate low <-- weak client ...:e3b8 signal -84 rate low <-- weak client ...:12aa signal -79 rate low <-- weak
Read it as a distribution, not as individuals. A healthy AP shows most clients in comfortable territory. Signal values are in dBm — negative numbers, closer to zero meaning stronger. Somewhere around −65 is a common comfort line, and beyond about −75 is increasingly grim. An AP where a third of the clients sit below −75, hour after hour, is telling you plainly what is happening. People regularly use the network from somewhere this AP barely reaches. The AP is saying, "There is a hole near me." Cross-reference with the complaint map and you can often name the corner before you ever stand in it.
Respect the caveats, because this data is honest but lopsided. It is the AP's side of the conversation, and the AP hears the client's small radio worse than the client hears the AP. So the data errs pessimistic. It only includes devices that managed to connect at all: a true dead zone produces no readings, just absence. And a weak reading might be a phone in a pocket in an elevator, not a desk. So treat AP-reported client signal as a tripwire and a pointer, and confirm with your own feet — which is the next section.
Because this data flows constantly, it can be promoted from a one-time look into a standing check. Compute, per AP, the share of associated clients sitting below your "grim" line, sampled a few times an hour during business hours, and trend it. Most APs will idle in the single digits. An AP whose weak-client share lives at a third and climbs after the furniture reshuffle is raising its hand on your behalf. That turns dead-zone detection from an annual project into a quiet background process. It is the same baseline-and-outlier thinking used across the AP fleet guide, applied to coverage.
The walk survey, step by step
A walk survey is exactly what it sounds like: move through the building with a device that records wireless readings. And note what you see, where. Professional tools exist that pair readings with a floor plan automatically and draw pretty heatmaps; they are lovely and entirely optional. A laptop or phone with any signal-scanning utility, plus a printed floor plan and a pen, produces a map you can act on. What matters is method:
- Step 1. Print the floor plan and mark a grid of survey points — every office or two, each conference room, corridor junctions, and every spot from the complaint map. Number the points.
- Step 2. Pick one device and stick with it. Different radios hear differently, so the whole survey must be one instrument. Use the scanning tool's view that shows signal per AP, ideally per band. Measure the band your users actually use, and note that the lower band reaches farther. So a spot can be fine on one band and dead on another.
- Step 3. At each point, stand still, hold the device consistently — same height, same orientation, not pressed against your body (you are made of signal-absorbing water). Give the reading a few seconds to settle, and write the strongest AP's signal next to the point number. Note the AP name too; "which AP serves this spot" becomes useful later.
- Step 4. Walk it twice if you can: once quiet (off-hours — pure coverage, no crowd effects) and once busy. A spot that is fine empty and awful when full is a congestion problem wearing a coverage costume.
- Step 5. Spot-check with real traffic at the marginal points. Signal is the map's backbone, but at any yellow or orange point it is worth thirty extra seconds: run a short ping to the gateway and load a page. A spot can show tolerable signal yet perform badly (interference, a struggling AP), and only an actual traffic test catches that. Signal draws the map; a traffic sample verifies the suspicious spots.
- Step 6. Note the weird. Big metal things, dense storage, aquariums, the server cage, freshly moved partitions — radio obstacles worth an annotation.
Tip: consistency beats precision. A survey done casually with one laptop, the same way at every point, is far more useful than a mix of readings from three different phones held three different ways. You are drawing a comparative picture, and comparisons only work when the instrument does not change mid-survey.
Putting results on the floor plan
Now translate numbers into colors, because nobody — including you — reasons well from a table of negative numbers. A simple banding that matches common rules of thumb:
| Reading at the spot | Color it | Meaning |
|---|---|---|
| Stronger than −65 dBm | Green | Comfortable for everything, including voice and video |
| −65 to −75 dBm | Yellow | Works for browsing; real-time apps start to wobble |
| −75 to −85 dBm | Orange | Marginal — slow rates, retries, user grumbling |
| Weaker than −85 dBm | Red | Effectively a dead zone for real work |
Color each survey point, and the building's story appears in a way no spreadsheet can match. Green pools around each AP, with yellow at the edges and — where the complaints said — an orange-red blotch. Add the AP locations to the same drawing and patterns leap out. A red zone equidistant from two APs is a gap in the plan. A red zone right next to an AP is a broken or badly placed AP; orange along one wall says the APs are all hugging the opposite corridor. This drawing is the deliverable. It converts an argument ("the Wi-Fi there is fine!" "no it isn't!") into a shared fact.
Reading the map: holes, shadows, and illusions
Before reaching for hardware, classify what you found, because three different creatures show up on coverage maps and only one of them is fixed by more signal.
True holes are places genuinely beyond every AP's useful reach — the classic corner farthest from everything. More or closer signal fixes them. Shadows are blocked spots inside otherwise good coverage: the room behind the elevator shaft, the corner past the kitchen's wall of appliances, anywhere behind serious metal. Shadows do not yield to power increases — the obstacle eats whatever you add — they yield to angle: serving the spot from a different direction. Illusions are complaint clusters that survey green. Strong signal but everything slow is airtime congestion — the RF interference guide explains that disease. Fine standing still but failing on the move is a handoff problem — that is the roaming guide. And "connects but nothing works" with a beautiful signal is usually a join-path failure, DHCP being the classic — the experience-metrics guide's join-path section covers unmasking it. Adding an AP to fix an illusion spends money to fix the wrong thing, which is precisely what the survey exists to prevent.
One subtle creature deserves its own paragraph: the band-specific hole. The lower frequency band travels farther and slips through walls better than the higher bands, so the edge of coverage is in a different place for each. A far corner can be tolerable on the long-reach band and stone dead on the fast one. In that case, because many devices prefer the fast band, they experience the dead zone while your survey (done on the other band) shows yellow. The symptom pattern is devices in that area performing poorly or flip-flopping between bands. This is why Step 2 said to survey per band: two maps of the same floor. And the truth usually lives on the higher-band one.
Fixing the hole
First, a step that saves face: check that the serving AP is actually healthy. A dead zone that appeared recently — "it used to be fine over there" — is rarely a design flaw, because buildings do not redesign themselves. It is usually an AP that died, lost a radio, or dropped into a low-power crouch on an oversubscribed PoE circuit, instantly converting its whole cell into a hole. Confirm the nearby AP is up, its radios enabled, and its uptime boring before treating the symptom as a coverage problem. Long-standing holes, by contrast, are design: proceed down the list, cheapest first:
- Reorient or re-aim. Antennas do not radiate in a sphere; a typical stick antenna throws a doughnut of signal around itself, strong to the sides, weak off its ends. An AP mounted sideways, or a doughnut aimed down a hallway instead of into the offices, wastes its energy on drywall. Costs nothing to try.
- Reposition the AP. A few meters matters: out from behind the beam, off the metal cable tray, from the closet ceiling into the open area, from the corridor into the room people actually occupy. Central, high, and unobstructed wins.
- Raise transmit power — cautiously. The obvious knob, and the most oversold. More power extends the AP's voice but not the clients' voices back. So it mostly manufactures zones where devices hear a network they cannot usefully answer. Loud APs encourage exactly the clinging behavior the roaming guide warns about. Small nudges only, and re-walk after.
- Add an AP. The real fix for a real hole. Serve the area from a fresh direction. Remember the new radio needs a channel that fits the plan, and the install needs a switch port with power. Check the closet's headroom against the PoE budget guide before promising anyone a date. Where no cable can go, a wireless backhaul unit (an AP that relays traffic over the air) can paper over a hole at the cost of airtime and reliability. That is a workable last resort, not a plan.
Validating with the same walk
After the change, resist declaring victory from your desk. Re-walk the affected points — same device, same spots, same method — and write the new readings next to the old:
| Survey point | Before | After | Verdict |
|---|---|---|---|
| P14 (NE corner office) | −86 dBm | −64 dBm | Fixed — green |
| P15 (NE window desk) | −83 dBm | −67 dBm | Fixed — yellow, acceptable |
| P9 (copy room) | −77 dBm | −76 dBm | Unchanged — shadow, needs angle not power |
The before/after table does three jobs. It proves the fix to you. It proves the fix to whoever approved the spending. And it catches side effects. A repositioned or re-powered AP changes the shape of coverage around it. So also spot-check a few previously green points nearby. And confirm the complaint stream from the area actually dries up over the following weeks. The complaint map was your first sensor; it stays your last one too.
For a fix you really need to trust — an executive corner, a newly wired production area — go one step further and leave an instrument behind. Use a spare laptop running a scheduled connect-and-test cycle from the once-dead spot for a week. That is the synthetic-probe idea from the experience metrics guide pointed at a specific patch of floor. And a week of clean joins and quiet pings is the strongest closure evidence a coverage ticket can get.
The living coverage map
You now hold a document that took real effort: floor plan, AP locations, colored survey points, walk dates and the device used. Do not let it die in a desk drawer. Buildings change — walls go up, storage racks fill, a department triples its headcount, someone installs a decorative metal wall. Every change edits your radio environment without filing a ticket. Keep the map where the next admin will find it. Note on it which device and method the readings came from (so the next survey is comparable). Re-walk affected areas after any renovation or move. The habits that keep such documents alive instead of instantly stale are their own craft, covered in the documentation guide. The coverage map is a textbook example of documentation that pays rent only while it is current.
And keep the always-on sources running between walks: the located-complaint tally and the per-AP weak-client distributions. Together they are a standing early-warning system that tells you when the map needs refreshing — which beats finding out from a slow leak of vibes.
Where to go next
The method in one breath: harvest complaint locations, mine AP-reported client signal, walk a consistent survey, color a floor plan, classify hole versus shadow versus illusion, fix cheapest-first, and prove it with the same walk. Dead zones are only one of the ways Wi-Fi disappoints people. The full experience picture, including the metrics that catch congestion and join failures, is in the experience metrics guide. The fleet-side data collection that feeds all of this is in the AP health guide. The whole series lives at Wi-Fi Monitoring: What Users Actually Experience.
Put this into practice with SysPeep Monitor
The walk is yours, but the always-on half of dead-zone hunting is a monitoring job. SysPeep Monitor discovers your APs and switches. It collects client counts and per-device health over SNMP. It keeps the dashboards and history that show which areas trend weak between surveys. The free trial sets up on a Windows machine in minutes.
Download Free TrialLearn About SysPeep