How mobile networks plan pop-up coverage for match days and concerts
Mobile coverage planning for events, from the outside in: which venues fill up, at what hour, how far the crowd spills, and how footfall helps place temporary cells.
Part of the guide Workflow automation
Mobile coverage planning for events starts with three questions: which venues in a city draw crowds big enough to strain the network, at what hour the crowd peaks, and how far around the venue it spreads. A network's own traffic counters answer that for sites it already serves well. Footfall from mobility data answers it from the outside, for every venue at once, so the radio team can decide where a temporary cell, an extra carrier or a staffed shop is worth the trouble.
This is for radio planners, event operations teams and retail managers at mobile operators and tower companies. It explains what footfall can tell a radio team and what it cannot, how to read a venue's week, a made-up season in Mumbai, the mistakes that put a cell on wheels in the wrong car park, and how an AI employee in xMap Flows can keep a venue watch running.
What footfall adds to your own network data
Your counters show the load on cells you have. They are weaker at telling you about a venue served by a rival site, a new arena that opened last season, or the streets outside a stadium where queues form before the gates open. Footfall measures how busy each small patch of the map is from a panel of devices scaled to the whole population, whoever their operator.
- It is people-hours, not people. Someone who stays three hours counts three times. That suits radio planning, since a long stay means long sessions on the network.
- It arrives about 10 days late. You read last month's match, not tonight's. Use it to plan the next event at the same venue, not to react during one.
- It starts on 17 October 2026. The record is short for now. Keep each week's figures yourself and the comparison grows with every event.
- It covers most of the world, not yet Australia. Check a venue's figures say data before you rely on them.
Reading a venue's week
Point Mobility Footfall at each venue and at a few spots around it: the main gate, the nearest station exit, the largest car park. Ask for footfall, busy hours and busiest days over the last 7 or 14 days. Then read three things:
- The busiest day against the quietest. A venue whose busiest day is several times its quietest is an event venue; one that is busy every day is a shopping or transit hub, which a permanent site should already serve.
- The peak hour and how the day splits. Evening concerts and afternoon matches load the network at different times, and so does the walk out afterwards.
- The spots around the venue against the venue itself. When the station exit is almost as busy as the stand, the extra capacity belongs between them, not inside.
A made-up season in Mumbai
This example is invented to show the method, with made-up figures. An operator's planning team in Mumbai watches five venues: a cricket stadium, an arena, an exhibition centre and two open grounds used for concerts. Each Monday, the flow reads the last 7 days at each venue and at its nearest station exit.
| Venue | Busiest day | Peak hour | Busiest against quietest day | Station exit against venue |
|---|---|---|---|---|
| Cricket stadium | Sunday | 16:00 | About 6 times | Nearly as busy |
| Arena | Saturday | 21:00 | About 4 times | Half as busy |
| Exhibition centre | Thursday | 12:00 | About 2 times | A third as busy |
| Open ground, north | Saturday | 22:00 | About 8 times | Twice as busy |
| Open ground, south | Wednesday | 19:00 | Under 2 times | Similar |
Invented figures, for illustration only.
The team reads it this way. The stadium and the north ground swing hardest, so they get temporary cells for the next big event. At the north ground the station exit is busier than the ground itself, because people queue and wait there after the show, so the cell goes near the exit. The exhibition centre's midday peak is steady enough that a permanent upgrade makes more sense than a pop-up. The south ground barely moves: the event there was small, and nothing changes.
Mistakes that put the cell in the wrong place
- Reading footfall as subscriber counts. It is presence across the whole population, not your customers and not devices on your network.
- Planning tonight's event from it. The data is about 10 days behind. It tells you how the last event went.
- Measuring only the stand. Queues, drop-off points and station exits carry much of the load before and after.
- Comparing an event week with any week. Compare the week with the event against a week without one at the same venue.
- Expecting figures in Australia. Footfall is not available there yet.
A venue watch as an AI employee
In xMap Flows this becomes one AI employee for the network team. The wider idea, AI employees that keep watching places for you, is on the location intelligence automation page. The blocks, in order:
- Schedule: every Monday at 08:00.
- My data with your venues and the spots around them, or Nearby places to list stadiums, concert halls, convention centres and event venues across the city.
- Mobility Footfall on each incoming place: last 7 days, with footfall, busy hours, busiest days and trend.
- Sort and limit by daily footfall, keeping the 15 busiest spots.
- xMap LLM on the whole list, asked to name the venues whose week swung most and where around each the crowd sat.
- Make a report or Email to the radio planning team.
What lands each Monday: a short note naming the venues that swung most, with each spot's busiest day, peak hour and daily footfall, and a PDF the team can attach to a request for a temporary site. Add a Google Sheets step to keep every week's figures, and in a month you have a venue record of your own.
Questions
Can footfall show how many of our customers were at an event?
No. It measures people-hours of presence across the whole population, not operator customers or devices on a network. Use your own counters for that.
How soon after an event can we read it?
Each day's footfall arrives about 10 days later, so you read an event about two weeks after it took place.
Can we pick a single event date?
The block reads a window of recent days, such as the last 7. Run it every week and keep the results, then compare the week with the event against one without.
Does it work in every country?
Footfall covers most of the world from 17 October 2026, but not yet Australia. A place with no data says so and costs nothing.