How no shows in audited attendance change the number an auditor signs

The board pack said 20,000. The certificate says 15,200. The finance lead wants to know where 4,800 people went, and the honest answer is that they were never anywhere, because no shows in audited attendance are handled by a rule that is one word long and easy to miss.

The rule sits in the middle of a bullet list that most organisers skim, and once you see it the whole shape of the audited number makes sense.

What does the standard do with a registration that never arrived?

UFI's Calculation Standards and Definitions, Annex 2 to the auditing rules dated June 2021, list what goes into the calculation of visits. Persons who purchase their tickets at the event. Persons who purchased their ticket before the event and who attend the event. Persons with a complimentary invitation or a reduced tariff ticket and who attend the event. Persons attending with a season ticket.

Three of those four clauses carry an attendance condition, and the standard capitalises the conjunction to make sure nobody misses it. Buying is insufficient. Being invited is insufficient. The evidence chain starts at admission.

So the auditor is not making a judgement about absence. Absence never enters the process. There is no population of no shows for the auditor to assess, discount or argue about, because the only records with standing are those tied to an entry during declared opening hours.

That has a consequence worth stating for anyone in finance. Revenue on a paid show is recognised on tickets sold and attendance is certified on tickets used. Those are different populations and they will never reconcile, and the difference between them is not an error in either system.

The gap is invisible on the certificate

Take the arithmetic plainly. Twenty thousand live registrations at doors, a 24 per cent no show rate, and 20,000 multiplied by 0.76 gives 15,200 verified visitors.

Now look at what an outsider can see. The UFI Standard Audit Certificate, the version UFI publishes in 2026 and marks 08/21, has fields for visitors, repeat visits, total visits, exhibitors and net exhibition space, with national and international splits. It has no field for registrations. The 20,000 appears nowhere. The 4,800 appears nowhere. The document shows the survivor and says nothing about the population it survived from.

Two things follow.

The first is that nobody can compute your no show rate from your certificate, and you cannot compute a competitor's from theirs. Any industry comparison of no show rates is built on self-reported numbers from organisers who each defined the denominator differently, and the definition of the rate itself is the whole argument before you even get to comparability.

The second is that an audited series will move for reasons your registration dashboard cannot explain. A show that recruits 12 per cent more registrations and suffers a 6-point worsening in turnout publishes a certified figure that barely moves, and the post-show conversation will be about a flat show when the marketing performance was strong and the turnout performance was poor. Keeping both series and reporting them separately is the only way that conversation stays useful.

Why does a blended no show rate move without anyone changing behaviour?

Because the blended figure is an average across segments with very different turnout, so it tracks your registration mix as much as your audience.

Work it through on the same show. Of the 20,000 registrations, 6,000 were paid delegates at 250 each and 14,000 were free trade visitors. Paid registrants turn up: assume 9 per cent absent, so 540 of the 6,000. Free registrants are less committed: assume 30 per cent absent, so 4,200 of the 14,000.

Total absence is 540 plus 4,200, or 4,740, which against 20,000 is 23.7 per cent. Close enough to the headline 24 per cent to be the same number.

Now change nothing about behaviour and grow the free audience. Next edition the paid group holds at 6,000 and the free group reaches 18,000, with both segment rates unchanged. Absence is 540 plus 5,400, or 5,940, against 24,000 registrations, which is 24.75 per cent. The blended rate has worsened by a point and every underlying group behaved identically.

The number that actually tells you something is the rate per registration type, tracked separately, with the mix reported alongside. A single blended figure will spend most of its life reporting the success of your free acquisition campaign as a decline in audience quality.

Where absence quietly costs you the international label

This is the case where no shows do real damage to something you can lose, and it catches organisers who monitor the wrong number.

UFI's internal rules set the international criterion at either at least 10 per cent direct foreign and multinational exhibitors, or at least 5 per cent foreign visits or visitors. The visitor route is measured on the audited figure, which means on arrivals.

Say your registration file looks comfortable: 1,300 international registrations out of 20,000, which is 6.5 per cent, a full point and a half above the threshold. Overseas registrants have to book flights, and their absence rate runs higher than domestic. Put domestic absence at 23 per cent and international at 38 per cent.

Domestic arrivals are 18,700 multiplied by 0.77, or 14,399. International arrivals are 1,300 multiplied by 0.62, or 806. Total arrivals are 15,205, and the international share is 806 divided by 15,205, or 5.30 per cent. Still above the line.

Now let international absence reach 45 per cent, which is one bad week for visas or one airline dispute. International arrivals fall to 715. Total arrivals are 15,114 and the international share is 4.73 per cent. The label criterion has been missed on a number that looked safe at registration, and nothing about the show changed except who managed to travel.

The mitigation is to monitor the international share on projected arrivals rather than on registrations, using last edition's segment turnout as the projection, and to check it while there is still time to recruit. The alternative route through exhibitors is also worth knowing, since a show near the visitor line may sit comfortably above the 10 per cent exhibitor threshold and can qualify there instead.

There is a related wrinkle in how the international figure can be established at all. UFI's rules require each international visitor to be registered in order to justify the number and percentage, and where a system of individual registration is genuinely impossible, UFI can authorise the organiser, on request and solely for international approval, to determine the total percentage of foreign visits through a sample carried out by a market research company and certified by an independent audit company, with the certificate required to state the sampling methodology. A sample drawn on the floor is taken among people who turned up, so it never carries the registration-side illusion described above. That makes it the more honest instrument for this specific question and the harder one to run, which is roughly the trade in every measurement decision here.

What to do before fieldwork

Three things, all cheap, all easier before the auditor arrives.

Snapshot the registration file at the moment doors open and keep the file. Once the show closes, records get cleaned, merged and deleted, and the denominator you would have used stops being reconstructable. That snapshot belongs in the evidence pack you build per edition alongside the extracts the auditor will ask for.

Tag registrations created on site distinctly. Somebody who registered at the door and walked in was never exposed to the risk of absence, so their presence in a turnout calculation flatters it by roughly the size of your on-site desk.

Separate cancellations from silent absence in your own reporting. The auditor treats them identically, because neither produced an admission, and you should not, because a cancellation is a person who told you and a silent absence is a person who did not. They have different causes and different fixes.

Where this stops

Nothing in the audit framework rewards you for reducing absence, and the certified figure improves at exactly the same rate whether the extra bodies come from better turnout or from more registrations. An organiser optimising only for the audited number will always find recruitment cheaper than retention, and the audited number will not tell them they made the audience worse.

The framework also cannot distinguish absence from a failure to record presence. Fifteen thousand two hundred certified visitors is consistent with 4,800 people staying home and it is equally consistent with 4,200 staying home and 600 walking through an entrance nobody was scanning. Those are opposite problems, one for marketing and one for operations, and the certificate reads the same either way. Separating them needs your own door-coverage evidence, which is part of the wider registered against verified decomposition.

One last honest limit. Absence is not evenly distributed across badge types, and the types with the highest absence are usually the free and invited ones, which are also the ones whose classification you agreed with the auditor in advance. Change the classification and you change the measured rate without changing a single person's behaviour.

This week, take last edition's file and compute turnout separately for your three or four largest registration types, then recompute the blended rate using this edition's projected mix. If those two numbers differ by more than a point, your headline rate is already reporting mix, and the definitions behind the audited figure are the place to start writing down what you actually mean by it.