Skip to content

Measuring the meeting no show rate without flattering yourself

MatchmakingUpdated 2026-08-188 min read

In short

The meeting no show rate is meetings scheduled minus meetings attended, divided by meetings scheduled. Cancellations withdrawn before a defined notice threshold leave the denominator; anything later stays in it and counts as a no-show. Split the result by side, because buyers and exhibitors fail in different proportions.

Two people report the meeting no show rate for the same show in the same week. One says 9 per cent. The other says 31 per cent. Neither of them made an error.

The first divided meetings marked as missed by every meeting that ever existed in the system, including the 640 proposals that were withdrawn in the weeks before the show. The second divided the same numerator by the meetings that were still live on the morning of day one. Same show, same data, two definitions, and the gap between them is larger than any improvement either of them could deliver in a year.

Three counts that get called the same thing

Before any rate is computed, name the counts. Most meeting systems hold at least four states and most reports collapse them into two.

Proposed. The engine suggested it. Nobody has agreed to anything, and these have no business appearing in a no-show calculation at all.

Confirmed. Both parties accepted. This is a promise with no time attached, and confirmations can be withdrawn.

Scheduled. Confirmed, placed in a slot, at a table, and still live at a defined cut-off. This is the denominator you want.

Attended. Both parties present. Whether one party present counts as attendance is a real decision, and I would say it does not, because a meeting where the exhibitor turned up and the buyer did not produced nothing except a staffed table.

The definition to publish is scheduled minus attended, over scheduled, with the cut-off written into it. Everything else is a variant and the variants can be reported alongside, once, so people stop reinventing them.

The reason this ordering matters is that the four states have different owners and different incentives. Proposals are the matching team's output, so a rate computed over proposals makes the matching team look excellent and tells operations nothing. Confirmations are the concierge team's output. Scheduled meetings are the only count that both teams and the exhibitor have all committed to, which is why it is the fair denominator, and it is also the count that maps onto a table sitting empty at a specific time.

Where should the cancellation cut-off go?

This is the only judgement call in the formula, and it deserves five minutes of thought rather than a default.

A cancellation two weeks out is capacity returned. The table gets reallocated, the buyer's slot gets refilled from the standby list, and nobody experienced anything. Counting that as a no-show punishes the participant who did the right thing.

A cancellation at 08:50 for a 09:00 meeting is an empty chair. The exhibitor staffed the table, the standby list cannot be worked in ten minutes, and the operational outcome is identical to somebody simply not arriving.

Set the threshold at the point where your own team can actually refill a slot. If your standby process needs 24 hours, the threshold is 24 hours. If you can send a push notification and rebook in 90 minutes, the threshold is 90 minutes and your no-show rate will look better than a competitor's for a real reason. Publish the number alongside the rate so nobody has to guess.

Event organisers already price this distinction when money is involved. IBTM World's hosted buyer terms for the 2026 edition commit each buyer to a minimum of 30 pre-scheduled or self-scheduled appointments and state that a fee of 100 euros might be charged for each missed appointment which has not been rescheduled. The rescheduling clause is doing the same work as a notice threshold: a meeting moved is not a meeting missed.

Working the arithmetic on one programme

Take a programme that ends the campaign with 2,400 meetings confirmed.

Between confirmation and the 24 hour cut-off, 310 are cancelled with notice. Those leave the denominator, so scheduled is 2,090.

On the day, 168 meetings have one party absent and 44 have both absent. Total meetings not held: 212. The no show rate is 212 over 2,090, which is 10.1 per cent.

Now split by side. Of the 168 single-party absences, 119 are the buyer and 49 are the exhibitor. Add the 44 double absences to neither side, or to both, and say which you did. Buyer-side no-show is 119 plus 44, which is 163, over 2,090, or 7.8 per cent. Exhibitor-side is 49 plus 44, which is 93, over 2,090, or 4.4 per cent. Those two do not sum to 10.1 because the 44 are counted twice, which is correct and needs a footnote.

The buyer figure being nearly twice the exhibitor figure is the finding. It points at travel, at competing sessions, and at buyers who accepted 30 appointments and could realistically hold 22. None of those are fixed by reminding exhibitors to staff their tables, which is what a blended 10.1 per cent tends to produce as an action.

What rate should you expect?

There is no published benchmark for trade show meeting no-shows that I would put in a board pack, and I would be suspicious of one that appeared without a stated denominator.

The nearest well-documented body of work is on appointment scheduling in healthcare, where the measurement problem is identical and the literature is large. Dantas, Fleck, Cyrino Oliveira and Hamacher reviewed it in Health Policy in 2018, selecting 105 studies from 727 retrieved, and reported an average no-show rate of the order of 23 per cent, ranging from 43.0 per cent in Africa down to 13.2 per cent in Oceania. Their review also found that the most commonly reported significant determinants of no-show were high lead time and prior no-show history.

Both of those transfer directly. Lead time on a hosted meeting confirmed in week 30 for a show in week 40 is ten weeks, which is long by any clinical standard. Prior no-show history is something you hold and almost certainly do not use, and a buyer who missed 4 of 12 meetings last edition is your best available predictor of what happens this time.

I would not quote the 23 per cent as an expected figure for a trade show. The populations are not comparable and the incentives are different. I would quote it as evidence that a no-show rate in the tens of per cent is a normal property of appointment systems rather than a sign your programme is broken, and that the interesting work is in the split rather than in the headline.

Counting attendance without lying to yourself

The rate is only as good as your attendance signal, and most programmes have a weak one.

A meeting marked attended because neither party complained records a default, and nobody measured anything. A meeting marked attended by the exhibitor's app tells you the exhibitor was there. The signal to want is both parties confirming, which nobody does reliably, so the practical options are a table-side scan of both badges, a two-question prompt sent to both sides at the end of the slot, or a sampled audit where a floor team physically checks 200 tables across the show and you weight the result up.

The sampled audit is the one I would build first, because it costs two people for two days and it calibrates everything else. If your system says 92 per cent attendance and a sample of 200 observed slots says 79 per cent, you now know your system's attendance flag is optimistic by 13 points and you can correct the whole series.

Sample size is worth a sentence, since somebody will ask whether 200 is enough. At a true attendance rate near 80 per cent, a simple random sample of 200 slots gives a standard error of about 2.8 percentage points, so the interval runs roughly 74 to 85. That is precise enough to catch a 13 point bias and nowhere near precise enough to tell you whether this edition improved on last edition's 79 per cent. Use the audit to correct the system's figure, then use the corrected system figure for the trend.

Registration no-shows at the front door are a different measurement with a different denominator, and the trade show no show rate work owns that one. Do not blend them, and be careful with buyers who never arrived at the venue, because every meeting they held is a meeting missed and they will dominate your worst cases.

Where this stops

A no-show rate measures absence and says nothing about why.

The most useful splits are ones your data probably cannot support. A buyer who missed a meeting because a session overran is a programming problem. One who missed it because a better meeting came up is a matching problem. One who missed it because the table was in a corridor nobody could find is a floor plan problem. All three arrive in the system as the same event, and the only way to separate them is to ask within a few hours, when people still remember.

The second limit is that the rate is not comparable across shows, and it barely survives comparison across editions of the same show. Change the notice threshold, the attendance signal or the quota and the series breaks. Keep the definition frozen for at least three editions before anybody draws a trend line through it, and when you do change it, restate the history.

The third is that a low no-show rate can be bought by scheduling fewer meetings, which is why this number should never appear without the meetings-held count next to it. A programme that scheduled 1,200 meetings and held 1,150 looks better on this metric than one that scheduled 2,400 and held 2,090, and it delivered 940 fewer meetings.

Start by writing your denominator down in one sentence and circulating it to the three people who report this number. If any of them are using a different one, the policy work on late cancellations and the standby queue that refills a slot are both being evaluated against a figure nobody agrees on, and your matchmaking reporting has a definition problem before it has a performance problem.

Questions people ask about meeting no show rate

How do you calculate a meeting no show rate?
Take the meetings that were scheduled and still live at the notice cut-off, subtract the meetings both parties attended, and divide by the scheduled figure. Write the cut-off into the definition, because a cancellation two weeks out is a different event from one at nine in the morning and mixing them makes the rate meaningless.
Should cancellations count as no-shows?
Late ones should. A cancellation that arrives with enough notice to rebook the table frees capacity and belongs outside the denominator. One that arrives after the table can be refilled leaves an exhibitor sitting alone, which is the same outcome as a no-show, so it belongs inside. The threshold is your choice and must be published.
Why report the no show rate separately for buyers and exhibitors?
Because the two sides fail for different reasons and respond to different fixes. Buyers miss meetings through travel, overrunning sessions and competing invitations. Exhibitors miss them through stand staffing, a busy walk-up queue or a more attractive prospect arriving. A single blended rate hides which of the two you should be working on.

Related reading

All matchmaking articles