Location Case Study UK Transport Hub

Liverpool Street Station
London, April 2026

Liverpool Street is one of the busiest stations in the UK — over 100 million passenger journeys a year, the gateway to the City of London, and the eastern terminus of the Elizabeth Line. NetworkUX surveyed the immediate station area at two contrasting moments on consecutive days in April 2026: a quiet Wednesday morning before the commuter peak, then the Thursday evening rush. The contrast is stark — and it isn’t a special event. This is what “normal busy” looks like to a major UK mobile network.

Liverpool Street Station mobile performance survey at Thursday evening rush hour, 16:00–18:00 on 16 April 2026 — four-panel map showing Vodafone, EE, O2 and Three. All four panels show overwhelming red and orange pin density, with large clusters of red 'no-entry' icons marking timed-out connections.
Thursday 16:00–18:00 — red pins are under-1 Mbps measurements; the red “no-entry” icons are connections that timed out completely. A normal Thursday evening.
Location
Liverpool Street, London EC2
Quiet baseline
Wed 15 Apr 2026, 07:00–07:30
Busy comparison
Thu 16 Apr 2026, 16:00–18:00
Networks surveyed
Vodafone · EE · O2 · Three
Measurements captured
1,940 across both surveys
The headline

Same place. Same week. A different network.

On a quiet Wednesday morning, all four UK networks delivered solid, consistent mobile data around Liverpool Street — zero failed connections, every measurement in the comfortable 5+ Mbps band. By Thursday evening rush hour, three of the four networks had failure rates above 50%, EE delivered zero measurements above 5 Mbps, and Vodafone’s average dropped to 1 Mbps. No event triggered this. No festival, no concert, no derailment — just a normal Thursday in the City.

Vodafone
Download
5.32 1.24 Mbps −77%
Upload
8.37 0.78 Mbps −91%
Failed connections
0.0% 87.1% of pts +87.1 pts
EE
Download
7.33 0.86 Mbps −88%
Upload
9.62 0.83 Mbps −91%
Failed connections
0.0% 67.1% of pts +67.1 pts
O2
Download
8.94 2.89 Mbps −68%
Upload
7.94 2.51 Mbps −68%
Failed connections
0.0% 29.2% of pts +29.2 pts
Three
Download
8.07 0.76 Mbps −91%
Upload
8.42 1.83 Mbps −78%
Failed connections
0.0% 59.7% of pts +59.7 pts
Download & upload, side by side

From comfortable to barely-functional — in nine hours.

Average download and upload speeds per operator at Liverpool Street, quiet Wednesday morning vs busy Thursday evening. The bars on the right side of each chart aren’t just shorter — they’re a different category of network experience.

Average download speed
Average download speed per operator at Liverpool Street, quiet morning vs busy evening 0 2 4 6 8 10 12 5.3 1.2 Vodafone 7.3 0.9 EE 8.9 2.9 O2 8.1 0.8 Three Speed (Mbps)
Average upload speed
Average upload speed per operator at Liverpool Street, quiet morning vs busy evening 0 2 4 6 8 10 12 8.4 0.8 Vodafone 9.6 0.8 EE 7.9 2.5 O2 8.4 1.8 Three Speed (Mbps)

Every network averaged 6.8 Mbps or better in the quiet morning. By the evening rush, EE and Vodafone had collapsed to under 1 Mbps on average — with upload speeds of 0.78 and 0.83 Mbps respectively, well below the threshold that completes a card transaction or posts a photo.

Connection quality, distribution

What the network actually delivered, broken down.

Each row shows the percentage of measurements that fell into each connection-quality band, classified by combined upload+download throughput. Failed = timed out or under 1 Mbps; Slow = 1–5 Mbps; Usable = 5–15 Mbps; Fast = 15+ Mbps.

Vodafone
Quiet AM
100%
214 pts
Busy PM
87% 9%
139 pts
EE
Quiet AM
100%
208 pts
Busy PM
67% 33%
304 pts
O2
Quiet AM
100%
259 pts
Busy PM
29% 59% 12%
407 pts
Three
Quiet AM
99%
171 pts
Busy PM
60% 35%
238 pts

The quiet morning bars are entirely cyan: every measurement on every network landed in the comfortable 5–15 Mbps range. The busy evening bars are a different shape entirely — EE’s row is two-thirds red, Vodafone’s row is overwhelmingly red, and O2 (the best of the four) still spent more than a quarter of its measurements in the failure band.

The maps

Same survey routes. Different reality.

Four operator panels each, all on the same map of the immediate Liverpool Street Station area. Each pin is a single measurement, colour-coded by speed: green is fast, red is slow, the red “no-entry” icons are timed-out connections that never completed at all.

Quiet baseline Wednesday 15 April 2026 · 07:00–07:30 Pre-commute — clean networks across the board
Liverpool Street Station mobile performance survey at 07:00–07:30 Wed 15 April 2026 — all four UK networks (Vodafone, EE, O2, Three) show predominantly yellow and green pins indicating 5–15 Mbps and faster performance, with zero failures.
What the colours say: Yellow and green pins fill all four operator panels. Not a single failed connection, not a single measurement under 5 Mbps. This is what every UK mobile network is capable of when the cells aren’t congested.
Busy comparison Thursday 16 April 2026 · 16:00–18:00 Normal evening rush — no special event
Liverpool Street Station mobile performance survey at 16:00–18:00 Thu 16 April 2026 — all four UK networks show predominantly red and orange pins, with large clusters of red 'no-entry' icons indicating timed-out connections. Vodafone, EE and Three show the most severe degradation.
What the colours say: The red “no-entry” icons aren’t decoration — each one is a measurement that timed out completely. Vodafone and EE are a wall of red. O2 keeps some yellow alive. None of the four networks delivers a single 15+ Mbps measurement in the immediate station area at rush hour.
For comparison

A normal Thursday in London is harder on the networks than the Fleadh was on Wexford.

NetworkUX surveyed Wexford Town during the 2025 Fleadh Cheoil — roughly 600,000 visitors flooding a town of 23,000 over a single weekend. Even under that 26× population surge, no Irish network exceeded a 5.2% failure rate. Liverpool Street on a normal Thursday rush hour produced failure rates of 29–87% across the four UK networks.

Worst failure rate, Fleadh weekend
5.2%
Eir, Wexford, 26× population surge
Best failure rate, Liverpool St rush
29.2%
O2, normal Thursday evening
Worst failure rate, Liverpool St rush
87.1%
Vodafone, normal Thursday evening
What this means

Four takeaways for transport hubs, retailers and operators.

01

Capacity is the daily problem, not the occasional one

This isn’t event data. There was no festival, no concert, no emergency. The networks at Liverpool Street collapse like this every weekday at rush hour because the cells are routinely operating at the edge of their capacity, and the commuter peak pushes them over.

02

Network choice changes the experience entirely

At rush hour, the gap between best and worst network at Liverpool Street is enormous. O2 delivered nearly three times the average speed of EE, and a third of EE’s failure rate. For retailers, transport operators and venues, this is actionable: pre-event communications, payment SIMs and IoT devices should preference the network that actually works at your location and time.

03

You can’t plan around it without measuring it

Operator-published coverage maps do not capture this. They show signal presence, not capacity behaviour at peak. Only ground-truth, multi-network measurement at the times that matter will give a venue, council or business the data it needs to plan around the real network.

04

Resilient infrastructure isn’t evenly distributed

The Fleadh comparison shows that mobile networks can stay up under extreme load when capacity is well-provisioned. The Liverpool Street numbers suggest that, in dense UK urban transport hubs, that provisioning hasn’t kept pace with growth in everyday demand.

Need this kind of survey for your location?

Measure the network you’re actually relying on.

NetworkUX provides multi-operator, multi-period mobile performance surveys for transport hubs, retail destinations, venues and councils — with the data presented in a form you can take to network operators, regulators or your own board.