At four o'clock the front door can tell you that 402 of 480 registered guests checked in. It cannot tell you that the panel in room three had eleven people in it and the workshop next door had ninety and turned people away.
Those are different facts, and only one of them changes what you do next year.
Most events measure attendance once, at the building. That was a fair approximation when everybody sat in the same room and watched the same thing. It is a worse approximation every year. Average session attendance fell from 5.3 sessions per person in spring 2025 to 4.2 in spring 2026, according to the RainFocus 2026 Mid-Year State of Events Report published on 2 September 2026. Over the same period event teams reported hosting up to three times as many meetings at their events as the year before. People are still coming. They are spending the day differently once they are inside.
If the only number you hold is the door number, that shift is invisible to you.
What the door number hides
A single check-in at the entrance gives you one true thing: this person was here. It is genuinely useful and for a lot of events it is enough.
It stops being enough the moment your programme has choices in it. Three things go unanswered:
- Which sessions earned the room they were given. A talk in a 200-seat room with 30 people in it was not a failure of the talk. It was a failure of the room allocation, and you can only see that if you counted.
- Who was actually turned away. A full room does not report itself. Unless someone counted at the door, "we were at capacity" is a story the crew tell afterwards, not a number.
- Who attended the thing you need to prove they attended. For accredited content, licensing hours or a sponsor's contracted audience, "they were in the building" is not evidence.
None of that is a reason to scan every door in the venue by default. It is a reason to decide, deliberately, which rooms are worth counting.
Decide which sessions to count
Counting everything is the version that fails. It doubles your scanning kit, needs a steward on every doorway and turns a five-minute changeover into a queue in a corridor.
Pick the rooms where the number changes a decision:
Capacity-limited sessions. Workshops, roundtables, anything with a hard seat count or a fire limit. Here the count is not analytics, it is operational. You need to know at the door whether the next person gets in.
Sessions people have to book. If attendees signed up in advance, you already have a list, and the interesting figure is the gap between who booked and who turned up. That gap is usually larger than organisers expect and it is the number that lets you release seats next time.
Sessions that have to be evidenced. Accredited or certificated content, sponsor-funded sessions with a contracted minimum, anything a funder will ask about.
One or two ordinary sessions, as a sample. If you count nothing else, count a couple of typical talks so you have a baseline to compare the special ones against.
Everything else can run open. A room with plenty of seats and no obligation attached to it does not need a scanner in the doorway.
Make the second scan cheaper than the first
The reason session check-in gets abandoned is almost always the queue it creates. Ten minutes between sessions is not long, and a corridor is a worse place to queue than a foyer.
A few things keep it quick.
Reuse the pass they already have. The badge that got them through the front door should be the same thing that gets them into the room. A guest who has to find a second QR code in a second email will hold up the line while they look for it. One signed pass, scanned again, is the whole mechanism.
Scan on the way in, never on the way out. Exit counting sounds tidy and never survives contact with a room emptying at once.
Put the scanner where the door already narrows. People slow down at a doorway anyway. Standing a steward two metres inside the room, where the flow has already spread out, makes the scan the thing that causes the queue.
Let the scan be silent when it is fine. The steward needs to look up only when something is wrong. A green tick and a quiet chirp for the 95% means the exceptions are the only thing anyone reads.
Two scanning devices per room, not one. Sessions start at the same time across the whole venue, so the peak is sharper than the front door's and much shorter. Two people scanning for four minutes beats one person scanning for nine.
What to do when the room is full
The honest answer is that this happens, and the point of counting is to handle it well rather than to prevent it.
Agree the number before the day, with the venue, and write it on the run sheet. Fire capacity and comfortable capacity are not the same figure, and the crew on the door need to know which one they are working to.
Then decide in advance what happens at capacity plus one. Is there a repeat of the session later that they can be moved to. Is there an overflow room with a feed. Is the answer simply "sorry, this one is full, here is what else is on now". Any of those is fine. What is not fine is a steward improvising a policy in front of a queue while the session starts behind them.
If the session is repeated, the scan record is what lets you tell the person at the front of the queue that they are booked onto the four o'clock one, without paperwork.
The numbers worth pulling afterwards
Session check-in earns its keep in the week after the event, not on the day.
Four figures do most of the work:
- Booked against attended, per session. The size of that gap tells you how much to overbook next time and which topics people commit to but skip.
- Peak concurrent attendance per room. Compare it to the room's capacity and you have your room allocation for next year, evidenced.
- Sessions per attendee. The average number of sessions each person attended, and the spread. A programme where most people attend one thing and leave is a different event from one where most people attend four.
- Turn-aways. Only countable if the crew log them, so give them a one-tap way to do it. This is the number that justifies a bigger room.
Set those against the front-door figure and you can say something more useful than "we had 402 people". You can say which parts of the day they chose.
Keep it proportionate
A single-track conference does not need session check-in. Neither does a networking evening, a launch, or anything where the room is the event.
The moment your programme starts making people choose, though, the door number stops describing your event and starts describing your car park. Counting two or three rooms properly is a small amount of extra work on the day and it is the difference between planning next year from what happened and planning it from what you remember.
Start with the rooms that have a hard limit. Add the ones you have to evidence. Leave the rest open, and let the front door do what it is good at.
