Visitor Flow Analytics, Camera or Smartphone When You Cannot Hand Out Tags

An organizing office preparing a four-day trade fair wrote this in an inquiry about visitor flow analytics. The event opens only four days a year, so buying equipment and leaving it installed is hard, and handing a tag to every visitor is hard too. Last month this blog covered assets you cannot attach a tag to. People are harder than assets. You have to ask for consent, you have to get the thing back, and you have to ask people to install an app.

Inquiries that come in from exhibition halls, museums and large public facilities all ask one thing in common. Can cameras do it, can beacons do it, can an app do it? The three are not competitors. They are three tools whose answer changes when the conditions change. This article sets out four of those conditions and sorts out which tool fits where.

I. Three ways to track visitor flow when you cannot attach anything to people

The first is the camera. ORBRO's AI RTLS video location tracking recognizes the movement of people and vehicles from camera footage alone, with no tags, and plots it on a map in real time. It can take the feed from CCTV already installed, and people counting by zone, dwell-time heatmaps and entry counting all come out on the same screen.

The second is the smartphone. If visitors have installed the event app or the exhibition guide app, the app scans the beacons stuck to the hall's pillars and reports which zone they are in. The beacons used for Mobile Location Tracking are standalone devices that need neither power nor network, so even for a four-day event it ends with sticking them up and taking them down. Accuracy is about 5 to 10 m, at zone level.

The third is the visitor pass tag. A UWB tag goes into a lanyard visitor pass and is handed out, so each individual is captured precisely and a specific visitor's path can be followed. In exchange, the operation of handing out and collecting comes attached. That operational design was covered in last month's post on exhibition visitor flow.

All three sit side by side on ORBRO's exhibition visitor solution page. Why the company lists three instead of pushing one is the next section.

II. Four conditions decide

Condition	Camera AI RTLS	Smartphone + beacon	Visitor pass tag
Is identity needed	Anonymous objects and headcounts	Per app account	Can be linked to the pass number
Event duration	Favors permanent sites, needs camera installation and mapping	Favors temporary events, only beacons to stick up	Anywhere, if you have staff to hand out and collect
Existing cameras	Straight away if present, installation if not	Not relevant	Not relevant
App install required	Not needed	Required, only those who install are observed	Not needed
Location unit	Coordinates within the camera's field of view, plus zone tallies	Zone, about 5 to 10 m	UWB precision, individual tracking
Blind spots	Outside the camera's view, where crowds overlap	People who did not install the app	Not worn, not returned
Operational load	Camera mapping and region of interest setup	Mounting beacons and distributing the app	Handing out, collecting, sanitizing

Two questions do the sorting. Do you need "who," or is "how many, where, for how long" enough? And what can you ask of the visitor? The less there is to ask, the more it leans toward the camera, and the more you have to follow a specific person, the more it leans toward the tag, while an app that already exists lets the phone fill the space between.

III. Where camera people counting fits

Permanent exhibition halls and large public facilities, where cameras are already mounted and the people keep changing. One public agency asked whether a specific subject's position could be followed on video in a facility with dozens already installed, and one art museum asked, at the construction stage before opening, how many people per unit area it could see. Neither question asks anything of the visitor.

What the camera method handles is movement, not faces. It catches a person on screen as an object and turns them into coordinates, counts how many enter and leave a zone, and colors in the spots where people linger. Which booth draws a crowd and where a path breaks come out of this method alone.

The limits are just as clear. Outside the camera's view is the same as not existing, objects break apart when crowds overlap, and recognition wavers when the lighting changes. So designing the camera method starts from how many units go where and which zones they will cover. Whether existing CCTV can be used as it is gets checked first against the five criteria in last month's CCTV post.

Here is the judgment. If the job is watching flow without identity, the camera is the tool that asks the least of the visitor.

IV. Where the smartphone fits

Events where visitors end up installing an app anyway. At a trade fair that handles ticketing, entry and exhibit commentary through an app, location observation starts by putting an SDK into that app. Beacons go onto the booth pillars and come off when the event ends. This is the method that fits that organizing office's condition, that new cameras cannot go up for a four-day event.

One thing has to be written honestly. Only the people who installed the app and granted the permission are observed. So the number is observed sessions, not total visitors. Whether the sample holds up statistically is this method's verification item, and that gets checked first in a pilot run.

Here is the judgment. With an app it is the lightest tool, and without one it does not hold up at all.

V. Where the visitor pass tag fits

When you need a specific visitor's position. Guide staff knowing an invited guest's path in advance, an alert when a child visitor leaves a designated zone, and counting precisely how many people are inside a demo area right now all belong here. The camera does not know "who" and the phone needs a favor asked, but the visitor pass is already put in their hand at entry.

The price of that is handing out and collecting. Tag quantity, breakaway lanyards for children, sanitizing and reuse, alerts for unreturned tags: all of that is the operational design, and it has to be settled before the positioning is.

VI. Why not do all three

There is one objection. Mount cameras, stick up beacons and hand out tags as well, and wouldn't that be the most accurate? Some sites are in fact built that way. Only one thing has to be understood going in. The three methods observe different subjects. The records of visitors who received a tag and the tallies of a stretch caught on camera are not counting the same people. Unless you decide in advance, by zone and by time slot, which of the two numbers to read as the reference when they disagree, two sets of accurate data end up casting doubt on each other.

So the order we recommend is one method per question. Congestion and dwell time on the camera, the paths of app users on the phone, a specific visitor on the tag. Even when all three land on a single ORBRO OS map, the question each one answers is different.

VII. What to decide before rollout

1. What you want to know, in one sentence. Whether it is "how many, where, for how long" or "who is where."
2. Duration. Permanent or a few days is what splits the camera from the beacon.
3. The position and specification of the existing cameras. Whether the feed can be used, whether the field of view covers the zones of interest.
4. What you can ask of the visitor. Installing an app, wearing a pass, or nothing at all.
5. The scope of the data you keep. Anonymous coordinates and zone tallies only, or linked to the pass number.
6. A pilot run. Check before the main event whether the sample holds up statistically and whether the camera count and placement are enough.

Closing thoughts

Look again at that trade fair office's conditions. Four days, equipment they cannot buy, tags they cannot hand out. Put the three conditions into the table and the answer narrows. With an app, beacons, and without one, a few temporarily mounted cameras and a region of interest setup. Tags for the invited guests only.

People are harder to attach things to than assets are. So design starts not with what to ask for, but with deciding what you do not have to ask for. ORBRO builds camera-based video location tracking, beacon-based mobile location tracking and UWB tags alike and puts them on one ORBRO OS map, so you can start with the single visitor flow method that fits your site conditions and add more when you need to. Send us the event duration, the space and the question you want answered, and we will sort out which method fits, together with you.