What an onsite rebooking rate actually measures and what it hides

Someone from commercial puts a slide up on the Monday after teardown. Onsite rebooking rate, 63 per cent. The room nods. Then the finance lead asks whether that is up or down on the last edition, and whether 63 is good, and it turns out nobody can answer either question without going back to the raw file.

The numerator is rarely the problem. Four hundred and two companies signed something during show week, and a signature has a date on it and a name against it. The denominator is the problem. Somebody chose it three editions ago, wrote it down nowhere, and it has been drifting ever since.

Which denominator is the rate actually using?

Take one industrial show with 640 exhibiting companies on the final floorplan. Here are the denominators I have seen used for the same 402 signings, on the same show, by people who all thought they were computing the standard measure.

Every exhibiting company present. 402 over 640 is 62.8 per cent.

Every company the sales team had a target on. Sales excluded 40 accounts they had already written off before the doors opened, so 402 over 600 is 67.0 per cent.

Every stand, rather than every company. A parent with three separate brand stands counts three times. The stand count was 712, so 402 signings over 712 is 56.5 per cent, and it is not even measuring the same thing, because one of those 402 contracts covered all three of that parent's stands.

Every company that could have rebooked at all. Strip out the ones with no next edition to buy and the base falls to 582, which puts 402 signings at 69.1 per cent.

That spread, 56.5 to 69.1, comes from one show, one week, one set of contracts. Nothing was manipulated. Four honest people picked four honest bases.

The version I would publish, and the second version next to it

Publish two numbers, always together, and name them so they cannot be confused in a deck.

Gross onsite rebooking is next edition contracts signed on site divided by distinct exhibiting companies present at the current edition. On the 640 base, 402 divided by 640 is 62.8 per cent. It is the simplest thing to compute, it is reproducible from the floorplan and the contract log without any judgement calls, and every operator in the business can check it.

Rebookable onsite rebooking uses a smaller base. Remove companies that had no ability to sign in the hall, and be strict about what qualifies. On our 640 that came to 58 accounts: 19 already under a multi-edition contract running through the next show, 22 co-exhibitors on somebody else's stand with no contract of their own, 11 government pavilion participants whose space is bought by a trade body, and 6 companies in administration or acquired mid-cycle. 640 minus 58 is 582. 402 over 582 is 69.1 per cent.

The gap between 62.8 and 69.1 is 6.3 points, and those 6.3 points are the part of your floor that was never available to your sales team on show week. Publishing only the gross figure makes your team look worse than they are. Publishing only the rebookable figure gives them a base they control the size of, which is an incentive nobody should be handed. Both numbers, with the exclusion list beside them, are the minimum any renewal intelligence reporting should carry on this measure.

The exclusion list is where the number gets gamed

Every exclusion needs a rule that a second analyst can apply to the same file and get the same 58.

The multi-edition contract exclusion is safe. It is a fact in the contract table with a start date and an end date, and it removes accounts that were structurally incapable of signing anything in the hall. The co-exhibitor exclusion is safe if your registration system distinguishes contract holders from listed co-exhibitors. If it does not, you are guessing, and the guess will be made by whoever runs the query.

The exclusion I would refuse is "not a realistic target". Sales teams will ask for it, and the argument sounds reasonable, because there genuinely are exhibitors nobody expected to rebook. Once a subjective category enters the denominator, the rate becomes a negotiation. If an account was not worth an onsite conversation, that is a fact about your show, and burying it in an exclusion list stops you from seeing how large that group has grown.

Can you benchmark this against an industry average?

Somebody in the room will have an industry average to hand. Before it goes on a slide, ask whoever is quoting it for the denominator, for whether it counts companies or stands, and for whether it measures a rebooking event during show week or a return appearance one edition later. The retention averages circulating on trade sites arrive with none of that, and the ones I have chased trace back through a chain of blog posts to a source that either does not resolve or never stated a definition in the first place.

Compare that with what CEIR does publish with its methods visible. The CEIR Index, reported through IAEE in 2025, tracks four separate components for the same event population, and in the second quarter of 2025 those four sat at different distances from their 2019 baseline: attendees down 3.7 per cent, net square feet down 4.9, exhibitors down 8.8, and real revenues down 15.6. Four measures of the same shows, a spread of nearly twelve points. If the industry's own index needs four numbers to describe one quarter, a single retention percentage detached from its definition is not going to settle an argument in your boardroom.

So my honest advice on benchmarking is to stop. Compare your show to itself across editions, holding the definition fixed, and compare shows inside your own portfolio only after you have confirmed they compute the base the same way. An external benchmark whose denominator you cannot inspect tells you nothing about whether 62.8 is good.

Deciding what counts as signed

The numerator has one genuine ambiguity, and it is worth settling in writing before show week rather than during the post-show review.

Three things happen in a rebooking booth. A company signs a contract for a named space at a stated rate. A company signs a space request or intent form with no space allocated and no price agreed. A company tells the rep they are in and the rep logs it in the CRM. Only the first is a contract. The second is a lead with a signature on it. The third is a note.

Count the first. Report the second separately as signed intent, because it converts at a rate you can measure and it is genuinely useful for pacing, and never add it to the contract count in a board pack. The rate at which intent forms become contracts belongs to the rebooking conversion funnel, which is its own topic. A contract signed in the hall is not money either, and the share of those 402 that cancel before the first payment falls due is a second figure that belongs on the same page as this one.

The parent company problem

The 640 in our denominator is distinct exhibiting companies, and getting to that number is harder than it sounds. A registration and contract file at a mid-sized show will carry the same corporate group under four or five legal entities, national subsidiaries, and a distributor who exhibits their brand.

Decide once whether you count legal entities or parents, then apply it to both sides of the fraction. Counting parents in the denominator and contracts in the numerator is a mistake that flatters you, because one parent contract can cover several entity-level stands, so the numerator stays whole while the denominator shrinks. On our show the entity count was 640 and the parent count was 601. Same 402 contracts, and the rate moves from 62.8 to 66.9 per cent purely on the choice of entity.

I would count parents, because a parent is the thing that makes a budget decision, and report the entity count alongside it so the floorplan team can reconcile. Whichever you choose, write it into the definition and restate it on every slide.

Where this stops

An onsite rebooking rate is a measurement of one week. It cannot tell you about the show it describes, because much of a show's renewal happens in the eight weeks afterwards, and a show that runs a hard onsite push will book a higher share in the hall and a lower share later without changing its final retention at all.

That makes the rate genuinely misleading for comparing two shows in a portfolio that sell differently. A show with a staffed rebooking suite, a floorplan release on day two and a discount that expires at teardown will post a high onsite rate by design. A show whose exhibitors need to see lead scan volumes before committing will post a low one and land in the same place by November.

The rate is still worth measuring, for two reasons. It is available on the Friday rather than in the following spring, and the share of your final book that closes in the hall is a real operating characteristic worth tracking across editions. Read it as a pacing measure and it earns its place. Read it as retention and it will mislead you every time the sales model changes, which is why an exhibitor retention rate definition is written separately and measured on a different date.

Pull your last edition's contract log and floorplan this week and compute both versions on one page: signings over exhibiting companies present, and signings over rebookable companies, with the exclusion list itemised by reason and count. If the exclusion list has a category in it that a second analyst could not rebuild from the data, that category is the first thing to fix.