A predictive survey models WiFi coverage on a computer using floor plans and assumed building materials. An on-site survey measures the real building with real equipment. Predictive work is quick and useful for planning and budgets; on-site measurements are what confirm a design will actually work, and most good projects use both.
How a predictive survey works
The designer imports scaled floor plans into planning software and draws in walls, doors, windows, elevator shafts and other features, assigning each one a material such as drywall, concrete block, brick or glass. The software applies a typical signal loss for each material, then lets the designer place virtual access points and see the predicted coverage, signal-to-noise ratio and channel overlap.
The result looks like a heat map from a real survey, which is both its strength and its risk. It is easy to present and discuss, but it shows what the model assumes rather than what the building does.
Where predictive surveys work well
- New construction, where there is no building to measure yet.
- Early budgeting, to estimate how many access points and cable drops a project needs.
- Simple, well-documented buildings with standard materials, such as a typical office fit-out with drywall partitions.
- Comparing options, for example ceiling-mounted versus wall-mounted access points, before committing.
Where predictive surveys fall short
- Unknown materials. Older buildings often hide wire lath plaster, metal-backed insulation, foil vapour barriers or extra layers of concrete behind finishes that look ordinary.
- Contents. Warehouse racking, stock, machinery, filing cabinets and even large crowds of people change how signals travel, and they change over time.
- Interference. A model cannot know about neighbouring networks, non-WiFi interference or the radio behaviour of a nearby tenant.
- Installation reality. Access points end up a few metres from where they were drawn because of beams, ducts, lights or sprinkler lines.
How an on-site survey works
A surveyor walks the building with survey software and measurement adapters, taking readings as they go. At ehZee we use AirMagnet Survey Pro. Depending on the question, the survey may be:
- Passive: listening to every access point and network in range, to map signal, SNR, channels and neighbouring networks.
- Active: connecting to the network to measure real data rates, throughput, packet loss and roaming.
- AP-on-a-stick: placing a temporary access point at proposed locations and measuring how far its signal reaches through the actual walls.
The on-site visit is also when the surveyor checks ceiling types, mounting options, cabling routes and the space in IDF and MDF rooms, none of which show up in a model.
Side by side
| Question | Predictive | On-site |
|---|---|---|
| Does the building need to exist? | No | Yes |
| Accounts for real materials and contents | Only as assumed | Yes, as measured |
| Finds interference and neighbouring networks | No | Yes |
| Checks mounting, ceilings and cabling | From drawings only | Yes |
| Useful for early budgets | Yes | Yes, if the building is available |
| Proves a finished network meets targets | No | Yes |
Combining the two
A common, sensible sequence on a project looks like this:
- Agree on requirements: users, devices, applications and the signal and SNR targets for each area.
- Build a predictive model to estimate access point count, locations and cabling.
- Visit the site to check materials, ceilings and the radio environment, and use AP-on-a-stick measurements in representative or difficult areas.
- Adjust the model and design based on what was measured.
- After installation, run a validation survey to confirm the network meets the targets, and tune it where it does not.
The predictive model saves time and gives everyone a shared picture; the on-site work catches what the model could not know.
Validation is not optional. Whatever approach is used for design, a survey after installation is the only way to confirm the network works as intended. Without it, a predictive design is an untested assumption.
How ehZee can help
We use predictive modelling where it saves time and on-site measurement with AirMagnet Survey Pro where accuracy matters, and we always validate after installation. Our WiFi surveys service can assess an existing network or support a new design, and our WiFi design and build service carries the work through to a verified network. If you are new to the topic, start with What is a WiFi site survey?
Common questions
Is a predictive survey enough for a small office?
For a simple space with standard construction, a predictive design may be close. A short validation survey after installation is still worth doing to catch surprises.
Can a predictive survey be done for a building under construction?
Yes, and that is one of its main uses. Plan for an on-site check once walls are up and before ceilings close, so access point locations and cabling can still change easily.
Why do predictive and measured results differ?
Usually because of materials, contents or interference the model did not include. The measured results reflect the real building and should take priority.