Vending machine location foot traffic: where people stay long enough to buy
Why time spent nearby predicts machine sales better than people walking past, how to read people-hours for a vending site, and how to plan restocking from busy hours.
Part of the guide Location intelligence automation
When judging vending machine location foot traffic, look for places where people linger, not just places they pass: waiting areas, hospital corridors, campus courtyards, platforms, laundrettes, gyms, workshops on a night shift. A machine sells when someone has a minute, a few coins and nothing better nearby. Footfall measured in people-hours suits this well, because a person who waits twenty minutes adds more than one who walks straight through. Rank candidate spots by footfall per day, then check when the busy hours fall and whether a shop or cafe a few steps away already serves the same need.
This guide is for vending operators, micro-market and smart fridge companies, and the facilities managers who host them. It explains why lingering matters, what to read in the data, a made-up example, the mistakes that put machines in the wrong corner, and a flow that both scores new sites and plans restocking.
Why lingering beats passing
Mobility Footfall counts presence: every hour someone spends near a point adds one people-hour. That is a different thing from a count of people walking by. A station passage with heavy, fast-moving crowds and a hospital outpatient wing with fewer but much longer stays can show similar daily figures for opposite reasons. For a poster, the passage might win. For a machine, the wing usually does, because waiting turns into buying.
The data cannot tell you which of these is happening, so you supply that judgement from the kind of place. Pair each candidate's footfall with a short note on its type: a waiting area, a thoroughfare, a workplace, a place people spend leisure time. Then compare like with like.
What to read for a vending site
| Figure | What it tells a vending operator |
|---|---|
| footfall_daily | How much presence there is overall, for ranking sites against each other |
| footfall_by_part | Whether the site lives in the day, the evening or late at night; a strong late night share is valuable where shops are shut |
| footfall_peak_hour | When stock runs down fastest, which sets the restocking window |
| busiest_day and quietest_day | Which days need a full machine and which can wait for the route |
| footfall_trend | Whether a site is getting busier or quieter week on week |
All in the place's own local time.
Keep the radius at its smallest. Up to 200 m the block reads the point's own hexagon of the map, about 0.1 km², which is the finest it goes and already takes in the building and a little around it. A machine serves only the room or corridor it stands in, so going wider just mixes in the street outside, a car park or the block next door.
Competition sits a few steps away
A busy lobby with a convenience store at the door sells fewer snacks from a machine than a quieter one with nothing nearby. Nearby places counts convenience stores, cafes and supermarkets within a short distance of each candidate from xMap's places database, and can pass on each site with a count rather than the list. Late at night that competition shrinks as shops close, which is why a strong late-night share and few rivals together make a good machine site.
A made-up example: four sites in Toronto
The figures below are invented to show the reasoning; they are not measurements. An operator has one new machine and four sites offered.
| Site | Footfall a day | Late night share | Busiest hour | Shops within 150 m |
|---|---|---|---|---|
| Office tower lobby | 18,000 | 3% | 12:00 | 5 |
| Hospital outpatient wing | 9,500 | 14% | 10:00 | 1 |
| Student residence | 6,000 | 29% | 22:00 | 0 |
| Subway passage | 31,000 | 8% | 17:00 | 4 |
Invented figures for illustration only.
On totals the subway passage wins easily, but people move through it fast and four shops sit nearby. The office lobby empties at night and is surrounded by places to eat. The hospital wing combines long waits with almost no alternatives. The student residence has the smallest total yet a third of its presence late at night and no shop at all. The operator places the machine in the hospital wing and puts the residence first in line for the next one.
Mistakes to avoid
- Ranking by total alone. High numbers in a thoroughfare often mean little time to stop.
- A wide circle. At 500 m you are measuring the neighbourhood, not the corridor.
- Ignoring the shop at the door. Count the competition within a short walk.
- Treating presence as sales. Use your telemetry or cash counts for that; use footfall to choose and to time.
- Expecting yesterday's figures. Plan from the pattern, which changes slowly, not from last night.
Site scoring and restocking with an AI employee
Two small flows in xMap Flows cover the job. The first scores sites offered by hosts; the second plans the restocking round. Both fit the idea of AI employees for field operations.
Scoring candidate sites
- Start with My data, the sites hosts have offered, each with its kind of place.
- Add Mobility Footfall: Footfall, Busy hours and Busiest days, 100 m, the last 30 days.
- Add Nearby places: convenience stores and cafes within 150 m, passing on each site with a count.
- Add xMap LLM: score each site from 0 to 100 for a snack machine, weighing lingering, late-night presence and nearby shops, with one reason.
- Finish with Sort and limit and Email to the sales team.
Planning the restocking round
A weekly flow brings in your installed machines, reads each one's busiest day and hour, and passes them to Route planner, which orders the stops from your depot. In the inbox: the week's round with each machine's peak day, so the busiest are filled before they empty. The busy stations with no convenience store template is a close starting point, and busy places with the fewest cash machines follows the same logic for another kind of machine.
Questions
Is a busier location always better for a vending machine?
No. Places where people wait or work for long stretches tend to sell more than busy thoroughfares, especially when no shop is close by.
What radius should I use for a vending site?
The smallest, 200 m or less, which reads the point's own hexagon of about 0.1 km². Two machines in the same building will share that figure, so tell them apart by the kind of space each stands in.
Can footfall tell me how many people use my machine?
No. It measures presence in the area. Your machine's own sales records show use.
Can I check sites in Sydney or Melbourne?
Not with footfall yet, which does not cover Australia. Nearby places still counts the shops around each site.