Trade versus consumer show counting rules and where the evidence diverges

A group runs a home and interiors event twice a year. The spring edition is trade only, badge and business card at the door. The autumn edition opens to the public on a ticket. Same halls, same organiser, same operations team, and the two post-show reports have almost nothing in common.

The trade versus consumer show counting rules differ because the evidence differs, and the evidence differs at the front door. A ticket proves an admission happened. A badge proves a named person was there. Each model produces a number the other model finds difficult, and most disputes about attendance figures are really disputes about which of the two chains somebody was standing in.

Which evidence chain does your show actually have?

Work backwards from what the till and the scanner produce.

At a consumer show, the primary record is a transaction. Somebody paid, and at the gate a barcode was checked and an admission incremented. The count of admissions is close to exact, because every one of them corresponds to a scan, a turnstile pulse or a torn stub. What the transaction does not carry is a person. Two tickets bought on one card might be one household or two strangers, and a ticket bought online and never used looks identical in the sales report to one that was.

At a trade show, the primary record is a registration. A named individual completed a form, was accepted, and collected a badge. The identity is strong, so unique people is the natural unit and repeat visits are a straightforward derived figure. What the registration does not carry is proof of payment or of qualification, both of which have to be established some other way.

UFI's Auditing Rules for the Statistics of UFI Approved Events, dated June 2021, accommodate both, and the accommodation is visible in the wording. Annex 2 includes in the calculation of visits people who purchase their tickets at the event, people who purchased before the event and attended, people with a complimentary invitation or a reduced tariff ticket who attended, and people attending with a season ticket. Every clause in that list describes consumer ticketing. The same annex then applies the visitor rule on top: a person is counted once for the whole run, and at most one visit per person per day.

What the standard says about a public show

The rules do not treat public shows as a lesser case. They give them a specific provision.

For an event to qualify as a UFI Approved International Event, either at least 10 per cent of exhibitors must be foreign or multinational, or at least 5 per cent of visits or visitors must be foreign. For public fairs, the rules state that this percentage is counted on the basis of professional visits or visitors, if they are identified. Annex 2 repeats the point on the visitor side: when the exhibition is a public show, it is acceptable that only foreign professionals are counted and compared against total professional visits, and "the auditor's certificate must identify the method of calculation used" (UFI, 2021).

There is a hard precondition underneath all of it. The UFI Internal Rules require that, for an audit to be valid, the organiser uses a registration system with entry controls. A consumer show selling anonymous cash tickets at the door with no gate scan has no chain to audit at all, whatever its turnstile counter says.

The UFI Approved Event programme itself certifies the same three quantities in either model: visitors split international and national, exhibitors split international and national, and net exhibition space, with visits countable alongside visitors when the certificate indicates it (UFI, 2026).

One weekend and one trade week, counted

Put real arithmetic on both chains.

The consumer edition runs three days. It sells 41,600 tickets, of which 33,200 online in advance and 8,400 at the gate. The gate scans 38,900 admissions, so 2,700 tickets were sold and never redeemed, which is 6.5 per cent of sales. Of the 38,900 admissions, 6,800 were second or third day entries by season ticket holders, so unique people is 32,100. That last figure exists only because the season tickets were named. Had they been bearer tickets, the honest report would stop at 38,900 admissions and record unique visitors as not measured.

The trade edition runs three days in the same halls. It takes 18,000 registrations and issues 14,200 badges at collection, a shortfall of 3,800 or 21.1 per cent, which is a pacing and no show question that P9 handles. Because badges are named, the 14,200 is directly a unique visitor count. Repeat entries across the three days add 5,900 controlled visits, giving 20,100 total visits and a visit factor of 1.42.

Line the two up and the asymmetry is clear. The consumer edition knows its admissions to the nearest scan and has to work to reach unique people. The trade edition knows its unique people from the badge file and has to work to reach a defensible visit count. Publishing 38,900 against 14,200 as though they were the same measurement compares an admissions total with a person count, and the ratio between them says more about ticketing than about demand.

Where the two models fail differently

Each chain has a characteristic failure, and knowing which one you own tells you where to spend audit preparation time.

The consumer show's failure is double counting. Without named tickets there is no way to distinguish a returning visitor from a new one, so a three day event with strong return behaviour can report a visits figure that a reader will hear as people. The correction requires either named ticketing, a survey based return rate certified by the auditor, or an honest statement that unique visitors is unavailable.

The trade show's failure is qualification. A badge proves someone came, and it proves nothing about whether they were the buyer the exhibitor paid to meet. The registration form's job title field is self reported, the approval workflow catches only what it was told to catch, and the gap between a verified attendance figure and a qualified one is the subject of P6. An auditor will certify the count and will not certify the quality, which is a boundary worth stating to a sales team before they quote a number.

There is a sample question sitting on top of both. The main published series for United States business to business exhibitions, the CEIR Index, is designed to be representative of the universe of business to business exhibitions in the United States (IAEE, 2026), which means a consumer show has no benchmark in it at all. Choosing a benchmark your show actually belongs to is P20's argument.

How do you count a hybrid of both?

Plenty of shows run two trade days and then open the doors, and the combined figure is where most misreporting happens.

Take a four day event. Days one and two are trade only, admitting 9,400 registered trade visitors with 3,100 repeat entries. Days three and four are public, selling 26,500 admissions. The temptation is to publish 39,000 as the headline, being 9,400 plus 3,100 plus 26,500.

That total is arithmetically fine and commercially misleading, because the exhibitor who bought a stand to meet specialist buyers cares about 9,400 and about nothing else in the sum. Report the two populations separately with their own bases, state the days each covers, and put the combined visits figure underneath if a combined figure is wanted. Where an exhibitor is on the floor for all four days, give them both numbers and let them price accordingly.

There is a leak between the two halves that catches people out. Trade visitors who miss the trade days buy a public ticket and come on day three, and they are then counted in the ticketed population with no role data attached. On the four day example, if 1,200 of the 26,500 public admissions were trade buyers, the trade audience for the week was 10,600 rather than 9,400, an understatement of 11.3 per cent on the figure exhibitors care most about. Recovering those 1,200 needs a question on the public ticket path asking whether the buyer is attending for business, which costs one form field and is the cheapest measurement improvement available to a hybrid show.

The residence rule for splitting either population into international and national follows its own logic, and P24 covers it.

Where this stops

Neither counting model produces a number that answers the question an exhibitor is really asking, which is how many people who could buy from them walked past the stand.

The consumer model gets closest on volume and furthest on relevance. The trade model gets closest on relevance and depends on a qualification process the organiser designed and the auditor will not assess. Running both models on one event, as a hybrid show does, gives you two partial answers and no way to combine them into one honest figure.

The other limit is structural. Where a consumer show has no named ticketing and no gate control, there is no audit available at any price, and the fix is a ticketing change that takes an edition to implement and affects the buying experience. That is a commercial decision rather than a measurement one, and it should be taken as such.

This week, write down which of the two chains produces your headline figure and what the last link in it is: a scan, a transaction, or a badge collection. Then check the published counting definitions for the base your chain can actually support, and whether the figure you published last time was one of them or a wider label that P13 would recognise.