How it works

From survey to evidence — in four phases.

NetworkUX is a measurement methodology, a delivery platform, and a way of doing operator conversations. Every survey follows the same four-phase process: define the area, capture every reading, visualise the result, and convert it into evidence you can act on. The pages below show that process in detail — with examples drawn from real surveys at festivals, stations, championships and parades.

4

UK networks measured simultaneously

100k+

Geolocated measurements per survey-day

2

Metrics captured: signal & performance

6

Audiences served with tailored reports

The methodology

Four phases. Every survey. Always the same.

The methodology doesn’t change between a community parade and a London transport hub. What changes is the survey area, the timing windows, the operators of interest and the audience the report serves.

01

Define the survey

Scope the area, the time windows, the routes.

Every NetworkUX engagement starts with a defined boundary: the borough, the BID footprint, the candidate charge-point sites, the venue and its approaches, the parade route. We agree the operators of interest (typically all four UK networks, or the three Irish networks for Ireland-based surveys), and the time windows that matter — quiet baseline, event peak, rush hour, festival night.

From the field At Liverpool Street Station, that meant a 200m radius around the station entrances, surveyed on a quiet Wednesday morning before commuter peak and again on a typical Thursday evening rush. Same routes, exactly one day apart — making the comparison airtight.

Liverpool Street Station quiet morning baseline survey — four operator panels showing solid usable speeds across the survey area.

02

Capture every reading

Foot, vehicle, fleet — whatever the route needs.

NetworkUX kit captures two streams simultaneously across every operator under test: signal (radio access technology — 5G, 5G NSA, 4G, 3G, 2G — and strength in dBm) and performance (download throughput, upload throughput and combined average in Mbps, with each measurement geolocated to within a few metres and timestamped to the second). Surveys are walked, cycled, driven or carried on a fleet vehicle for ongoing coverage. Every reading is recorded.

From the field At The Open Championship, that produced 22,075 signal measurements across four networks in a single day — including the surprising finding that Three’s 4G signal was present on 99% of measurements but 99.9% of throughput tests timed out completely.

Vodafone signal and performance map at The Open Championship 2025 — left panel shows signal strength, right panel shows download performance, both rendered as overlays on the Royal Portrush course.

03

Visualise the result

Colour-coded maps. Per-operator views. Before-and-after pairs.

Raw measurement files become readable products. RAG-coloured pins for performance, hex grids for coverage density, side-by-side operator panels for like-for-like comparison, pre-event and during-event pairs that show the impact of crowd load. Time-of-day filters, address-level lookups, and operator selection let stakeholders interrogate the same evidence at the level of detail they care about. Coverage grids render at H3 resolution 12 — 10-metre hexagons — finer than Ofcom Map Your Mobile (50–100 m) and the operators’ own published checkers. See the comparison →

From the field The Fleadh case study visualises the same Wexford Town survey routes for the Saturday before and the Saturday of the Fleadh, network-by-network — making the 26× population surge story visible at street level.

Manchester four-network coverage checker — interactive web map showing all four UK operators side-by-side with hex-coded performance bands.

04

Act on the evidence

Reports that hold up in an operator meeting.

The methodology only matters if the output drives a decision. Every NetworkUX engagement closes with a deliverable scoped to the audience — a council planning report, a CPO LEVI bid attachment, an operator-ready evidence pack, an event-day debrief for venue stakeholders. Underlying raw data is exportable in standard formats (KML, shapefile, GeoJSON, CSV, PDF) so the data lives wherever the customer’s workflow needs it.

From the field The St Patrick’s Day Parade case study handed the council a quiet-day baseline they can compare against next year, plus a defensible record showing that all three networks coped — with one operator’s service actually improving on parade day.

St Patrick's Day Parade in Wicklow Town — Vodafone parade-day performance map showing predominantly green and yellow pins along the Main Street route.

What we capture

Two metrics, two streams, every reading geolocated.

Every NetworkUX measurement lands as a row with these fields. They’re the vocabulary every case-study chart and report on this site is built from.

Signal

What the radio can see.

5GStandalone 5G — dBm strength
5G NSANon-standalone 5G (most current UK 5G) — dBm
4GLTE — dBm
3GUMTS / HSPA — dBm
2GGSM — dBm (rare on modern surveys)
No signalCell unreachable — count and locations recorded

Performance

What the network actually delivers.

AvgCombined upload + download average (Mbps)
Av DLAverage download speed (Mbps)
Av ULAverage upload speed (Mbps)
Min / MaxSlowest / fastest measurements in the sample
Bands% in 0–1 / 1–5 / 5–15 / 15+ Mbps tiers
Timed outTests that never completed at all

Every reading is also tagged with: precise lat/lon (sub-10m), timestamp to the second, network operator, device identifier, and the survey campaign metadata. That’s the structure that makes a NetworkUX dataset filterable by time-of-day, operator, location and event — not just a pile of speed tests.

Survey modes

The method adapts to the brief.

A NetworkUX survey can be run in any of four modes — or any combination.

Walked / cycled

For dense routes — parade lines, station forecourts, BID footprints, candidate charge-point sites. The right mode where pedestrian density and route specificity matter more than coverage breadth.

Vehicle-mounted

For wider areas — borough-scale coverage maps, intercity arterials, council concession boundaries. Captures the same data structure but at vehicle speed and proportionally wider footprint.

Fleet-equipped

Place a battery-powered NetworkUX kit in a council waste-collection truck, parking-enforcement vehicle or service fleet and capture coverage continuously while the vehicle does its normal work. The kits are portable, so they swap between vehicles each day to target specific routes — the lowest-cost route to ongoing borough-wide mapping.

Per-site validation

For candidate charge-point or IoT sites — a focused, repeatable survey of a small geographic point captured across every operator. The right mode where the question is “will this specific site work” rather than “how does the area look”.

What you get

Outputs your team can actually use.

Every engagement produces deliverables in formats that fit how your team works — whether that’s a written report for a council committee, a map a planner can pin to a wall, or raw data your GIS team can drop into their stack.

PDF reports

Audience-tailored written reports — council committee briefings, LEVI bid attachments, event-day debriefs, BID renewal evidence. Charts, maps and narrative integrated.

Interactive coverage maps

Web-hosted four-network maps your stakeholders can filter, zoom, and search by address — like the Manchester and Tees Valley examples we’ve published.

Raw data exports

KML, shapefile, GeoJSON and CSV exports for your GIS team or your operator’s radio planning team. Every column, every reading, every timestamp.

Operator-ready evidence packs

Pre-packaged datasets formatted for direct submission to an operator’s network planning function — the kind of evidence that turns a complaint into a conversation.

See it in action

The methodology, in real surveys.

Four recent NetworkUX engagements that exercise the full methodology — from event stress tests to daily commuter-peak collapse to small-town parades.

Pilot · Regulator · Rail

Ofcom Brighton–London Rail — 2025

44 journeys, 7 days, 4 MNOs — commissioned by Ofcom

Read case study

Event · Venues & Events

Fleadh Cheoil — Wexford 2025

5.2% worst failure rate, 26× population surge

Read case study

Event · Venues & Events

The Open Championship — Royal Portrush 2025

Signal everywhere, service nowhere — Three 99.9% failure

Read case study

Location · Mobile Infrastructure

Liverpool Street Station — London 2026

A normal Thursday: 87% failure rate at peak

Read case study

Event · Venues & Events

St Patrick’s Day Parade — Wicklow 2026

A network that actually got faster on parade day

Read case study

Programme · Local Authorities

Tees Valley Coverage Mapping — 2025

10M+ datapoints, 5 councils, 4 networks — via bin wagons

Read case study

Programme · Local Authorities · with video

Westminster × Veolia — 2024

1 week, 1.2M datapoints, 8 sq mi — the model that started it

Read case study

Programme · Local Authorities · major city

Manchester City Council × Biffa — 2025

4 weeks, 2M+ datapoints, every MCC street — bin-wagon model at city scale

Read case study

For your audience

Methodology stays the same. The report changes.

The four phases above are the same whichever audience the report is for — but the scoping, the analysis lens and the deliverable change to fit the customer’s decision context.