People counting for events: foot traffic, without a turnstile

If you need to know how many people are through the gate and how many are inside right now, you don't need a turnstile and you don't need a camera rig. People counting for events comes down to a pair of cheap wireless sensors either side of an entrance: they count who walks in and who walks out, keep a live occupancy figure, and send it to your phone. Here's the call I'll plant a flag on straight up — for the vast majority of events, a directional beam or radar sensor on a defined doorway beats every fancier option on cost, speed and hassle. A turnstile stops people; you only want to count them. Here's how each method actually works, what it really measures, and where the bargain gear lets you down.

What you're actually trying to measure

Before you buy anything, get clear on which number you want, because they're not the same and they need different gear. There are three:

Total attendance — how many people came through all day. A simple count-up. Useful for the wash-up, the sponsor report, and next year's planning.
Live occupancy — how many are inside right now. This is entries minus exits, running in real time. It's the one that matters on the day: capacity limits, when to open a second bar, when to hold the gate.
Flow and busyness — where people are and when it's heaving. A heat read across the site over time, not a precise headcount.

The mistake is buying a tool that gives you a trend when you needed a hard number, or a hard number at one door when you needed the whole-site picture. Pick the number first. Everything below follows from that.

The honest answer for a gate: a directional sensor

For entries, exits and live occupancy at a defined entrance, the reliable tool is a directional sensor watching the opening — a break-beam, a small radar, or a time-of-flight unit. The trick that makes it count direction, not just bodies, is two beams a few centimetres apart. Whichever beam breaks first tells the sensor which way the person walked. Break the outside one first, that's an entry; the inside one first, that's an exit. So it isn't tallying crossings blindly — it's keeping a running occupancy: in, out, and how many are inside at this second.

This is the same idea as the little counter above a shop doorway, sized and powered for an event instead of screwed to a wall. It counts a body. It never takes a photo, never needs decent light, and doesn't care who the person is — which is exactly why it's the low-drama option. The count goes back over a wireless link to something you can watch live, and that link is the part people forget to plan. We'll come to it.

Why a turnstile is usually the wrong tool

A turnstile feels like the "proper" way to count, and for most events it's a downgrade. It forces everyone single-file through a barrier, which is slow at a busy entrance, builds a queue you didn't need, and costs real money to hire and install. You're paying to stop people when all you wanted was to count them. A pair of sensors over the same opening counts the identical crossings, lets the crowd walk through at its own pace, and packs into a road case. Reach for a turnstile only when you genuinely have to physically gate entry — ticket validation, a hard capacity barrier — not when the real job is a number on a screen.

The phone-detection trick — useful, but not a headcount

You'll see "count the crowd by their phones" pitched as the clever modern way, and it has a real use, so let's be precise about it. Phones constantly broadcast WiFi probe requests and Bluetooth signals as they look for networks and devices. A sensor can sit and tally those, which gives you a genuinely handy read on busyness — which zone is heaving, when the afternoon lull hits, busy versus quiet across the whole site over time.

What it is not is a headcount, and anyone who tells you otherwise is selling. Three reasons it can't be:

Phones randomise their address. Modern iOS and Android deliberately rotate the identifier in those probes specifically to defeat this kind of tracking. One phone looks like many.
People carry more than one device. A phone and a watch and earbuds is three signals, one human.
Phones off the air vanish. Phone in a bag with WiFi off, or just out of range, and that person isn't counted at all.

So use phone-signal sensing for the relative, whole-site trend — and a directional gate sensor when you need an actual number on the door. They answer different questions. Don't let one masquerade as the other.

And cameras — only when you need what a camera sees

Camera-based counting is the heavyweight. Done well it counts and also gives you dwell time and rough demographics, which beam sensors can't. But it's the most expensive and most demanding option by a distance: it wants power, a clean line of sight, decent light, real processing grunt, and — the one people skip — a privacy story. The moment you point a recording camera at your patrons to count them, you've taken on questions about what you're capturing and storing that a beam sensor never raises, because a beam never sees a face. For pure entry, exit and occupancy, a sensor that counts a body without ever forming an image is cheaper, simpler and far less of a headache. Use a camera only when you specifically need dwell or demographics, not just a count.

The bit everyone underestimates: getting the numbers off the gate

Here's where cheap people counting quietly falls over, and it's never the sensor. The sensor counts fine. The problem is the link. A counter that knows your live occupancy but can't tell you is useless — and most events run on a showground, in a hall, or out in a paddock where there's no reliable fixed line to lean on. So plan the connectivity first:

Short, indoors, one or two gates: event WiFi back to a base unit is fine, as long as the WiFi itself is solid. A counter on flaky WiFi drops numbers exactly when it's busiest.
Spread out, multiple gates, big site: a long-range low-power radio link from each sensor to one base beats trying to blanket the whole site in WiFi. The sensors send tiny amounts of data, so they're ideal for a long-range link that reaches across a showground on a battery.
No fixed line at all: a 4G or 5G connection at the base is what carries the live count off-site to a phone or a cloud dashboard. At a venue with no NBN, the mobile network is your uplink, so check the signal at the actual gate position before the day, not from the car park.

This is the order I'd sort it in: connectivity, then counting, then the dashboard. Get it backwards — buy the clever counter, work out the link on the day — and you'll be standing at a gate with accurate numbers nobody can see. If you're working out how to get any signal onto a site with no fixed line in the first place, that's the foundation everything here sits on — start with reliable event WiFi and connectivity where there's no NBN, then layer the counting on top.

Make the count honest: lane the entrance

The single biggest thing that wrecks accuracy isn't the sensor's spec sheet — it's the doorway. A directional sensor on a sensible, defined width is comfortably 95 percent accurate or better, which is plenty for occupancy and attendance. It falls apart when people walk through three abreast, when a surge packs bodies over the beam at once, or when prams and trolleys and someone hauling a marquee pole confuse a simple counter. None of that is fixed by spending more on the box.

It's fixed by where you count. Funnel the entrance with barrier or bunting to roughly a metre and a half per sensor so people pass in countable lanes, and put the sensor where the crowd is squeezed, not where it fans out into the grounds. Count at the pinch point. A cheap sensor on a well-laned gate beats an expensive one on a wide-open paddock entrance every time. Same principle as everything else here: the gear is the easy part, the setup is where the number is won or lost.

The bottom line

People counting for events isn't a turnstile and it isn't a capital purchase. It's a pair of directional sensors on a properly laned gate, counting bodies in and out without a camera or a barrier, a connectivity link sized to the site so the numbers actually reach you, and someone watching the live occupancy on the day. Decide which number you need first — attendance, live occupancy, or busyness — then match the tool to it: beam or radar at the gate for a real count, phone-signal sensing for a whole-site trend, a camera only when you need what only a camera sees. Set up like that, "how many are inside right now?" stops being a guess at the gate and becomes a figure on your phone.

Frequently asked questions

What is people counting for events and do I need a turnstile?

People counting for events is measuring how many people enter and leave a space so you know your live occupancy and your total attendance. You do not need a turnstile. A turnstile forces everyone single-file through a gate and is slow, expensive and miserable at a busy entrance. A pair of cheap wireless sensors mounted either side of an opening counts the same crossings without a barrier, sends the count to your phone or a dashboard, and packs down in a road case. For most events the sensors are the right answer; a turnstile is only worth it when you genuinely have to stop people, not just count them.

How do wireless people counters actually count entries and exits?

The reliable method at a defined gate is a directional break-beam or a small radar or time-of-flight sensor that watches the opening. Two beams a few centimetres apart see which one breaks first, so the sensor knows whether someone walked in or out and keeps a running occupancy, not just a total. The count is sent back over a wireless link — WiFi, or a long-range low-power radio for spread-out sites — to a counter you can watch live. It is the same idea as a shop doorway counter, sized and powered for an event instead of bolted to a wall.

How accurate is wireless people counting at an event?

A directional sensor on a single, defined doorway is comfortably 95 percent or better, which is fine for occupancy and attendance. Accuracy drops when the entrance is wide and people walk through three abreast, when a crowd surges and bodies overlap the beam, or when prams, trolleys and people carrying gear confuse a simple counter. The fix is to count at a controlled width — lane the entrance to roughly a metre and a half per sensor — rather than to buy a more expensive box. Count where people funnel, not where they fan out.

Can I count a crowd by detecting their phones over WiFi or Bluetooth?

You can estimate a crowd that way, but treat it as a trend, not a count. Phones broadcast probe and Bluetooth signals a sensor can tally, which is genuinely useful for a whole-site footfall trend or a busy-versus-quiet heat read. It is not a headcount: modern phones randomise their address to defeat exactly this, one person may carry two devices, and someone with their phone off vanishes. Use phone-signal counting for relative busyness across a site over time, and a directional gate sensor when you need a real number on the door.

Do the sensors need internet or NBN on site to work?

The counting does not, but getting the numbers to you usually does. The sensor itself counts locally whether or not there is internet. To watch live occupancy on a phone or feed a dashboard you need a link off the gate — event WiFi, a 4G or 5G connection, or a long-range radio back to a base unit. At a showground or a paddock with no fixed line this is the part to plan first: a cheap counter is useless if its numbers are stuck on the sensor. Put the connectivity in before you worry about the dashboard.

Is camera-based people counting better than sensors?

A camera can count and also tell you dwell time and rough demographics, but it is the heaviest option: it needs power, a clean mounting position, good light, real processing and a privacy notice, and it raises questions you may not want to answer about recording your patrons. For pure entry, exit and occupancy at a gate, a beam or radar sensor counts a body without ever capturing an image, which is cheaper, simpler and far less of a privacy headache. Reach for a camera only when you specifically need what a camera uniquely sees, not just a number on the door.

How much does people counting for an event cost?

Far less than people expect, because the sensors are cheap commodity hardware. The real cost is in the setup: getting the connectivity onto the site, laning the entrances so the count is honest, and having someone watch the live occupancy on the day. A simple two-gate count on hired sensors is a modest day rate, not a capital purchase. The expensive mistake is buying turnstiles or a camera analytics rig for a job a pair of beam sensors does for a fraction of the price.

Running an event and want to actually know your numbers without a turnstile slowing the gate? That's what we do — work out which count you need, lane the entrances so it's honest, and get the connectivity onto the site so the live occupancy lands on your phone. No barriers you don't need, no analytics rig for a job a beam sensor does. Tell us about your event and we'll point you straight.