How many check-in lanes does an event need?

Most events are staffed by copying what was done last time. That works until the size changes, and then it fails in the only fifteen minutes anyone remembers.

This page is the arithmetic underneath our other door guides, written out in full so you can run it on your own event, disagree with any of our assumptions, and substitute your own numbers.

The formula

Lanes = spike arrivals ÷ (15 × the wait you will accept, in minutes), rounded up.

Four inputs, one of which is a judgement call rather than a measurement. Everything below is how to get each one.

Step 1: plan against attending, not registered

This is the step most people skip, and it is the one that costs money in both directions.

Registrations are not arrivals. A free-to-attend corporate event loses a meaningful share of its list on the morning, so sizing the door against the registration count buys lanes and crew for people who were never coming. Sizing it against a guess buys too few.

Take your registered number, apply your own no-show rate, and carry the result forward. If you do not have your own rate yet, the 500-pax runbook explains why published benchmarks disagree so widely and how two of your own events beat any of them.

Worked example. 700 registered, and your last two events ran about 30% no-show, so you plan for 490 attending. Round it to 500 and move on. Precision here is false comfort.

Step 2: find the arrival spike

The useful number is not the headcount. It is how many people walk in during the busiest fifteen minutes.

Corporate guests do not spread themselves across a registration window. They leave the office together, they share transport, and they aim to be seated just before the start. On our events roughly 60% of the room arrives inside one 15-minute band immediately before the programme begins.

Worked example. 500 attending × 60% = 300 arriving in the spike.

Three things push that share higher, and they are worth checking before you accept 60%:

• A hard start time. A Minister, a CEO, or a livestream that begins on the minute compresses everything.
• One shared arrival. Buses from a single office, or one car park with one barrier.
• A single entrance. The building becomes the constraint before your lanes do.

Two things push it lower: a genuine breakfast or networking hour that people actually use, and a staggered programme where the first session is optional.

Step 3: pick the wait you will accept

This is the only input that is a decision rather than an observation, and it is the one that changes the answer most.

A ten-minute worst-case queue is normal and nobody complains. A five-minute one costs twice the lanes and buys you a room that is seated before the client walks in.

Ten minutes is our default because it is the target that reproduces the staffing table in our crew-sizing guide exactly. Five minutes is what the 500-pax runbook uses, because that scenario has a Minister speaking at 9:15.

Step 4: divide

One lane clears about one guest every four seconds once it is warmed up — scan, confirm, hand over a badge, next. Call it 15 guests a minute.

So a lane absorbs 150 guests against a 10-minute target, or 75 against a 5-minute one.

Worked example, continued. 300 in the spike, hard start, 5-minute target:

300 ÷ (15 × 5) = 4 lanes.

Same event with a soft rolling start and a 10-minute target: 300 ÷ 150 = 2 lanes. The event did not change size. The promise you made about the queue did.

Worked examples in full

The last row is the useful one. Ten lanes is a real answer and it is usually the moment someone reconsiders the arrival plan instead — two entrances, or a genuinely staggered start, is cheaper than ten lanes and ten people.

The arithmetic is allowed to tell you the event is wrong. That is most of its value.

The walk-in desk is not a lane

Add one walk-in and exceptions desk on top of whatever the formula returns, and never count it in the lane number.

The moment a lane stops to solve a problem, everyone behind that problem stops too. One confused guest costs a lane two minutes, which at 15 a minute is thirty people who did not move. A lane that handles exceptions is not a lane, and the formula quietly stops being true.

Below about 100 in the spike the desk and the lane can be the same table with two people on it. Above that they should be physically apart, with separate signage, because guests self-sort by what they can see.

Where these numbers come from

Stated plainly, because a method you cannot audit is not a method.

Three of those four are ours to be wrong about. If you measure your own and they differ, your numbers win — the formula is the durable part, not the constants.

What this method does not cover

It sizes check-in throughput, and nothing else. These are the things that break it, all of which we have watched happen:

• Security or bag screening. A separate queue upstream, usually slower than check-in, and it sets the real arrival rate at your lanes. Size that first and the spike arriving at your door is whatever screening lets through.
• Badge printing on demand. A print station is not a lane. Printing, trimming and sleeving a badge runs well over four seconds, so on-demand printing needs its own stations and its own arithmetic. See badges and lanyards.
• Physical bottlenecks upstream. One lift, one escalator, one car park barrier. The door cannot run faster than the building feeds it, and adding lanes behind a constraint achieves nothing except a wider place to stand.
• Multiple entrances. Split the spike by entrance first, then run the formula once per entrance. Guests do not redistribute themselves to your shorter queue; they use the door they were dropped at.
• On-site payment or contract signing. Any transaction at the door is an exception by definition. It belongs at the desk, not in a lane.
• Events under about 50 people. The arithmetic returns 1 and so does common sense.

Measuring your own numbers

Each constant above is measurable at your next event, at no cost, if check-in writes timestamps to one record.

1. Lane rate. Take any five minutes at full flow, count the scans on one device, divide by five. Do it at the peak, not at the start.
2. Spike share. Plot your check-in timestamps in 5-minute buckets. The busiest three consecutive buckets are your spike. Divide by total attended.
3. No-show rate. Registered minus attended, over registered. Two events give you a usable figure and five give you a good one.
4. Wait target. Not measurable. Ask the client what a queue photograph costs them.

After one event you can replace our defaults with yours. That is the point of writing the method down rather than publishing a staffing table and hoping your event resembles ours.

The short version

Work out who is actually attending, take 60% of them as your fifteen-minute spike, decide whether you are promising a ten-minute queue or a five-minute one, and divide by 15 a minute. Add a walk-in desk that is never counted as a lane. Then check that nothing upstream — screening, one lift, one entrance — is going to make the whole calculation moot.