What a Wireless Site Survey Tells You Before You Buy
A wireless site survey tells you how many access points to order, which model, where each one goes, and how many cable drops and PoE ports sit behind them. It does that by measuring or modelling how signal actually behaves on your floor, instead of guessing from square metres. For a Dubai office, the survey is what turns "we need Wi-Fi on level 14" into a purchase order you will not have to redo once the ceiling is closed.
The catch is that "site survey" covers three quite different jobs, and a report from one of them cannot answer the questions the others answer. This article explains what each type produces, what a useful report contains, and the warning signs in a report that should stop you placing the order.
The three kinds of survey, and what each one can prove
Cisco's site survey guidelines (last updated 14 November 2023) describe three methods. A predictive survey is done in software from floor plans and wall types, with no field measurements. A passive survey walks the floor in listen only mode, recording every access point and signal the survey adapter hears. An active survey connects a client to the network and measures what that client experiences, including retransmissions and data rate changes.
Cisco Meraki's survey documentation adds a practical variant for the weeks before ordering: an "AP on a stick", where the exact access point model you intend to buy is mounted on a stand at the planned position and powered from a PoE injector, so you measure the real radio rather than a simulation of it.
| Survey type | Best moment in a UAE fit-out | What it tells you | What it cannot tell you |
|---|---|---|---|
| Predictive | When the fit-out drawings are approved and partitions are fixed | A budget count, rough placement, cable drop positions | Real wall losses, neighbouring networks, interference |
| AP on a stick | After partitions and glazing are up, before the ceiling closes | The true cell size of the model you will buy, room by room | How the finished network performs under load |
| Passive and active validation | After installation, with furniture in and people working | Whether the design targets were met, and where they were not | Anything you can still change cheaply |
The order matters. Cisco's guidelines say predictive work suits a space that "has not yet been built" and budgetary planning, while active surveys are recommended for new deployments. A predictive plan on its own is a starting estimate, and it is only as accurate as the wall materials someone typed into the software.
Why the walls decide the count in a Dubai building
Cisco's Wireless RF Reference Guide (updated 18 April 2024) gives typical losses of 3 dB for drywall, 10 dB for a brick wall and 12 dB for concrete. Every 3 dB halves the signal power, so the gap between a gypsum partition and a concrete core is roughly a factor of eight. A predictive model that treats a JLT tower's service core as drywall will place an access point that looks fine on screen and leaves the lift lobby and the rooms behind it with almost nothing.
Two materials common in UAE offices are not in that list and are the ones most often guessed: coated solar control glass on the facade, and metal framed glass partitions around meeting rooms. Neither has a safe default value to type into a model. This is the single best reason to put a real access point on a stick in those rooms before you commit: one measurement replaces a guess that decides how many units you buy.
Six GHz makes the walls matter more. Cisco's Wi-Fi 7 and 6 GHz migration guide (updated 22 July 2026) states that 6 GHz signals experience 2 dB more attenuation than 5 GHz and can be more affected by obstacles. That is why the 6 GHz band, which the UAE regulator TDRA opened for indoor Wi-Fi at 5925 to 6425 MHz in its announcement of 27 December 2020, needs its own measurement rather than an assumption carried over from 5 GHz.
The numbers a survey should design to
A report is only useful if it states the target it was measured against. The figures Cisco publishes are the common reference points:
- Signal at the cell edge. Cisco's 2023 survey guidelines and Meraki's documentation use -67 dBm for voice grade coverage. Cisco's 2024 RF Reference Guide notes that -66 dBm is now the better survey value for newer Wi-Fi phones.
- Signal to noise ratio. A minimum of 25 dB for voice, which Cisco expresses as a -67 dBm signal over a noise floor no higher than -92 dBm.
- Overlap for roaming. At least two access points heard at -67 dBm or better at any point where people walk on a call.
- Co-channel separation. Access points on the same channel should not hear each other above -82 dBm, per the RF Reference Guide.
If your office runs Teams or Webex calls from laptops and phones, design to the voice numbers even if nobody owns a Wi-Fi handset. A floor designed only for "coverage" at -75 dBm can show green on a heat map and still leave calls breaking up in the corridor between two meeting rooms, which is exactly where people walk while talking.
What a survey report should hand you before you order
Cisco's guidelines list the minimum contents of a survey report. Use it as a checklist when a report lands on your desk:
- Heat maps of signal strength at the stated target, for each band the access points will use.
- Heat maps of signal to noise ratio, not signal alone.
- Noise floor readings and spectrum analysis screenshots naming any interference source found.
- The exact access point hardware and antennas used during the survey.
- The walking path, so you can see which areas were actually measured.
- Rogue and neighbouring access points detected.
For purchasing, ask for one more page: a bill of materials that maps each access point on the plan to a model number, a mounting bracket, a cable drop and a switch port. That page is what lets a supplier quote accurately and lets you check the switch PoE budget before anything is delivered. Our guide to choosing a network switch for a Dubai office covers how to size that budget.
Warning signs that should pause the order
- The survey used a different access point. Meraki's documentation says surveys should use the same model that will be deployed. A cell measured with one vendor's antenna pattern does not transfer to another's.
- Only a signal heat map, no SNR. In a Business Bay or JLT tower, the tenants above, below and next door all run their own Wi-Fi, so a strong signal can still sit on a crowded channel. The passive survey is where those neighbours show up.
- No 6 GHz data for a Wi-Fi 6E or Wi-Fi 7 order. Cisco's July 2026 migration guide says a new 6 GHz project "must also include a dedicated 6 GHz site survey". The same guide treats one for one replacement of 5 GHz units as workable only where the existing access points already cover the floor while running at least three to four transmit power levels below maximum.
- Outdoor areas planned on 6 GHz. TDRA's 2020 designation is for indoor use, so a terrace, loading bay or yard in Jebel Ali should be designed on 5 GHz.
- The pantry was ignored. Cisco's VoWLAN troubleshooting guide names microwave ovens, Bluetooth devices and cordless phones as the most common 2.4 GHz interferers. Meeting rooms next to the kitchen deserve a spectrum check.
- A predictive plan presented as final. If the report never mentions a site visit, treat the count as a budget figure, not an order quantity.
What goes wrong when the survey is skipped
The expensive failure is not the access point. It is the cable. In a typical Dubai fit-out, data points are pulled while the ceiling grid is open, and adding a drop after the tiles go in means lifting them again and bringing a cabling contractor back to site. Buying four units, discovering the boardroom needs its own, and then finding no cable above it is the pattern a survey prevents.
The second failure is buying the wrong generation. A tri-band Wi-Fi 7 access point on a floor where the survey shows every enclosed room behind concrete may need a denser layout than the Wi-Fi 6 units it replaces, because of that extra 2 dB of loss at 6 GHz, which changes the cable count and the PoE budget. The trade-offs are covered in our Wi-Fi 6E versus Wi-Fi 7 buying guide, and the capacity side of the count in how many access points a Dubai office floor needs.
Turning the survey into a quote
Once the report meets the checklist above, the order is mostly transcription: access point model and quantity, mounts, licences if the platform needs them, and the switching and cabling behind them. Codeeo Tech supplies wireless hardware rather than carrying out surveys, and carries access points from Cisco and HPE Aruba. Every item is priced on request, so send the survey's bill of materials through the contact page and ask for a quotation against that exact list.
Cover photo: Office Drop Ceiling HVAC Vents and Lights by Tony Webster, via Wikimedia Commons (CC BY 2.0).
