First time exhibitor retention is where most of your churn lives

The new business report for our food and drink show closed the year with 143 first time exhibitors, which was a record and got a slide of its own. The following edition, 80 of them were gone.

Nobody had noticed, because the show's overall retention rate barely moved. One number covering 780 accounts absorbs a loss of 80 without visible strain, and the retention line went from 72 per cent to 71 and got explained away as noise. It was not noise. It was one cohort losing more than half of itself while every other cohort held.

Split the rate by tenure before you split it by anything else

Most organisers who break retention down break it by category first, because that is how the sales team is organised. Category tells you which halls are struggling. Tenure tells you where the mechanism is, and the mechanism is more actionable.

The split needs one field: editions completed, counted for each exhibiting company as of edition N. A first time exhibitor is a company with exactly one edition behind it, meaning edition N was its first. Then a second edition cohort, a third, and bands after that, because the counts get thin and the differences get small.

Two definitional points decide whether the number is reproducible. A company that exhibited five years ago, lapsed, and came back last edition is not a first time exhibitor and should not be counted as one, so the tenure field has to read the full history and not the last two rows. And a company appearing for the first time as a co-exhibitor on somebody else's stand is a first time exhibitor if you count co-exhibitors in your denominator at all, which is a decision your exhibitor retention rate definition should already have settled.

The arithmetic on one show

780 exhibiting companies at edition N. Split by editions completed, with how many of each cohort return at edition N plus 1.

First edition, 143 companies, 63 return, 44.1 per cent. Second edition, 118 companies, 78 return, 66.1 per cent. Third edition, 104 companies, 76 return, 73.1 per cent. Fourth to sixth edition, 218 companies, 172 return, 78.9 per cent. Seventh edition or more, 197 companies, 167 return, 84.8 per cent.

The cohorts sum to 780 and the returns sum to 556, so overall retention is 556 divided by 780, which is 71.3 per cent. That is the number on the board slide, and it is an average of rates running from 44 to 85.

Now count the losses instead of the survivors. The show lost 224 companies. 80 of them were first time exhibitors. So a cohort holding 143 of 780 accounts, which is 18.3 per cent of the base, produced 80 of 224 departures, which is 35.7 per cent of the churn. Strip the debutants out entirely and the remaining 637 companies retained 493, a rate of 77.4 per cent, six points above the headline. Which of those 80 are worth calling back a year later is a separate question with a different answer for each.

The comparison that settles where to spend is between two plausible improvements. Move first time retention from 44.1 to 60 per cent and the cohort returns 86 companies instead of 63, so total returns go to 579 and overall retention reaches 74.2 per cent, a gain of 2.9 points. Move the entire tenured base from 77.4 to 80 per cent, which is a real and difficult piece of work across 637 accounts, and returns go to 573, giving 73.5 per cent, a gain of 2.2 points.

A 16 point improvement on one small cohort outperforms a 2.6 point improvement on everybody else. That is the case for treating year one as its own programme with its own owner, and for pointing renewal intelligence at the cohort rather than at the average.

Why does the tenure curve rise at all?

The tenure curve invites a causal story, and the story is usually about onboarding. Year one exhibitors do not know how the show works, their stand is badly placed, nobody briefed them on lead capture, so they have a poor edition and leave. All of that happens and it is worth fixing.

It is not the only thing producing the shape. Fader and Hardie, writing in the Journal of Interactive Marketing in 2007, set out the shifted beta geometric model, in which every account carries its own fixed probability of renewing and those probabilities vary across the population. Their result is that the observed retention rate of a cohort increases with tenure even though no individual account ever becomes more loyal. The accounts with low renewal probabilities leave earliest, so the survivors at each successive edition are drawn from a progressively more committed slice of the original group.

Your first edition cohort contains every speculative exhibitor who booked a shell scheme to see what the show was like, alongside the companies that will still be with you in fifteen years. Your seventh edition cohort contains only the second kind, by construction. Some of the 41 point gap between 44 and 85 per cent is sorting rather than experience.

This matters because it caps what an onboarding programme can deliver. If half the gap is sorting, a first time retention rate of 60 per cent is ambitious and 75 is not available at any price. Fitting the model to three or four cohorts of your own gives you the split, and it does so with two parameters estimated in a spreadsheet, which is about the right amount of machinery for the amount of data a show actually has.

What is a debutant worth if you keep one?

The revenue argument for caring about year one is weaker than the count argument, and it is better to say so than to have a finance lead find it.

First time exhibitors take small stands. On our show they averaged about 4,200 of space against a show average near 9,800, so the 143 debutants held roughly 600,600 of a 7.6 million floor, which is 7.9 per cent of revenue from 18.3 per cent of companies. The 80 that left took about 336,000 with them, which is 4.4 per cent of exhibit revenue. Read on this edition alone, first time churn is a third of your departures and a twentieth of your money.

The case has to be made on expected life instead, and the tenure curve gives it to you directly. Chain the observed transition rates and a debutant survives to a second edition with probability 0.441, to a third at 0.441 times 0.661, which is 0.292, to a fourth at 0.213, to a fifth at 0.168, and from there at roughly 0.848 per transition. Sum the whole series and the expected number of editions for a first time exhibitor is about 3.05.

Now condition on surviving the first transition. An exhibitor present at its second edition has an expected 4.65 further editions ahead of it, so counting its first, about 5.65 in total. One retained transition roughly doubles the expected life of the account, and it does so at the point where the account is cheapest to influence, because you have a stand number, a named contact and a fresh set of scan data to talk about.

That is the honest version of the business case. Year one churn is not expensive this edition. It is expensive in the fourth edition that never happens.

The number the industry repeats, and what it is worth

The figure that circulates is 44 per cent, and it is worth knowing where it comes from before you benchmark against it. It is from the 2015 Benchmarks and Trends in Exhibit and Sponsorship Sales survey, a joint project of Exhibit Surveys, Lippman Connects and Trade Show Executive drawn from 206 respondents involved in exhibit and sponsorship sales, which put year to year exhibitor retention at 76 per cent and found that 44 per cent of first time exhibitors were satisfied enough to sign on for the next edition of their events.

Three things follow. The figure is self reported by sales people rather than computed from contract files, so it inherits whatever definition each respondent had in their head. It is from 2015 and now circulates on trade sites with the survey stripped off, credited to one of the three partners with no year and no denominator attached. And the gap it implies, 44 against 76, is a 32 point spread between debutants and the general population, which is close to the 33 point spread on our own show between 44.1 and 77.4.

Use it as an order of magnitude that says your first edition cohort should be losing roughly half its members, and treat anything better than 55 per cent as worth investigating for a definitional error before you celebrate it.

Industry level counts will not help you here either. The CEIR Index, produced by the Center for Exhibition Industry Research with Tourism Economics, an Oxford Economics company, measures the year on year change in four things, one of which is the number of exhibiting companies, and its 2026 report forecasts the total index growing 2.1 per cent in 2026. A net count of exhibiting companies nets debutants against departures, so a show can hold its exhibitor count flat for five years while losing every first time exhibitor it ever signs. Only the cohort split shows it.

Where this stops

The tenure split confounds with size and with category, and on a show this size the cells get small fast.

First time exhibitors are smaller than average, and smaller accounts churn more at every tenure. So part of the 44 per cent is a size effect wearing a tenure costume, and if you want to know how much, compare first time exhibitors against tenured exhibitors holding stands of the same size band. On our show that comparison is available for the 9 to 18 square metre band and nowhere else, because there are only eleven debutants above 36 square metres and a rate computed on eleven accounts is not a rate.

The deeper problem is that some share of your first edition cohort never intended to come back. A company launching one product, a distributor testing a market, a national pavilion participant whose slot was paid for by a trade body. These accounts are indistinguishable from genuine prospects in the contract file, they are counted in the denominator, and no onboarding programme reaches them. If you can capture stated intent at contract signature, even as a single field on the booking form, the cohort splits into a population you can work on and one you cannot, and every rate above becomes interpretable. Coding each departure against a fixed reason taxonomy does the same job after the fact, which is slower and still better than free text.

Take your exhibitor file for the last three editions, add a column counting editions completed as of each edition, and compute retention for the one edition cohort against everybody else. Two numbers, one join, and the gap between them tells you immediately whether your churn problem is a year one problem or a general one.