Net revenue retention for exhibitions when booth sizes move both ways

A finance director asked a question in a budget review that took the room twenty minutes to fail to answer. If we sold nothing new next edition, no first time exhibitors at all, what would exhibit revenue be?

Four people had a retention rate to hand. None of those rates answered the question, because retention rates count accounts surviving and the question was about money arriving, and between the two sits every exhibitor who came back and took a different amount of space than last time.

What question does the measure answer?

Fix a cohort at edition N: every exhibiting company present, with the space revenue each of them contracted. Go to edition N plus 1 and total what that same set of companies contracted, whatever they did. Some are gone and contribute zero. Some came back at the same size. Some took more. Some took less.

Net revenue retention is the second total divided by the first. Companies exhibiting for the first time at edition N plus 1 are excluded from both halves, because the whole point is the behaviour of a fixed group.

That exclusion is the discipline, and it is the part people skip. A show can add fifty new logos, post revenue growth, and be shrinking on every account it already had. The only number that reveals it is one that refuses to count the new business.

Gross revenue retention answers a narrower question. It weights each surviving account by what that account was worth in the base year and ignores what it bought this time, so it measures relationship survival with a size weighting and cannot exceed 100 per cent. Net revenue retention can exceed it, and whether it does turns out to be more interesting for an exhibition than for most businesses that use the measure.

The bridge on one show

512 exhibiting companies at edition N, contracting 6.80 million of space revenue between them, an average of 13,281 each.

372 of those companies return at edition N plus 1, which is 72.7 per cent of the count. Between them they had contracted 5.61 million at edition N, so the 140 companies that left held 1.19 million. The leavers averaged 8,500 each, the returners 15,081.

Gross revenue retention is 5.61 divided by 6.80, which is 82.5 per cent.

Now open up the 372 returning accounts and look at what they actually bought.

158 of them took the same space as last time, 2.44 million on both sides. 121 took more, moving from 2.05 million to 2.87 million, an increase of 0.82 million. 93 took less, moving from 1.12 million to 0.81 million, a decrease of 0.31 million. The three groups sum to 372 accounts and their base year revenue sums to 5.61 million, which is the check worth running before you go any further.

At edition N plus 1 those 372 accounts contract 2.44 plus 2.87 plus 0.81, which is 6.12 million. Net revenue retention is 6.12 divided by 6.80, which is 90.0 per cent.

The bridge from 100 reads as four numbers, each expressed as a share of the 6.80 million base. Churn takes 1.19 over 6.80, which is 17.5 points. Contraction takes 0.31 over 6.80, which is 4.6 points. Expansion adds 0.82 over 6.80, which is 12.1 points. Start at 100, subtract 17.5, subtract 4.6, add 12.1, and you land on 90.0.

Those four numbers are the whole measure, and each one has a different owner. Churn belongs to the renewal team. Contraction belongs to whoever handles the accounts that renewed. Expansion belongs to the person selling upgrades off the floorplan. Publishing all four is what turns a ratio into renewal intelligence somebody can act on.

Why does the software benchmark not transfer?

Net revenue retention arrived in general use through subscription software, and the benchmarks came with it. SaaS Capital's 2025 retention work on private B2B companies reports median net retention rising with average contract value, and for companies with contracts between 25,000 and 50,000 US dollars it puts the median at 102 per cent, the top quartile at 111 and the lowest quartile at 97. A show director who reads that and takes 100 per cent as a pass mark has imported a target that the physics of a hall will not deliver.

Here is why. A software account can add seats without any other account giving one up. Booth space does not work that way. If your hall was sold out at edition N, then every square metre a returning exhibitor adds at edition N plus 1 is a square metre that some other exhibitor released, either by leaving or by shrinking.

Work the ceiling. Hold the sellable area fixed and hold the rate card fixed. The maximum revenue the returning cohort can contract at edition N plus 1 is the whole hall, which is 6.80 million, the same total as edition N. So net revenue retention on a full hall with unchanged prices is capped at 100 per cent, and it only reaches 100 if returning exhibitors take every stand and you sell nothing at all to a new exhibitor.

That makes the ceiling a planning number rather than an aspiration. If your acquisition target is 12 per cent of the floor going to new exhibitors, your net revenue retention ceiling this cycle is 88 per cent. Our show came in at 90.0, which means returning accounts absorbed slightly more of the hall than the plan allowed for and the new business team got less room than they were promised. Read that way, net revenue retention stops being a retention number and becomes a mix number, which is a more honest description of what it measures on a capacity constrained show.

A show can clear 100 per cent. It needs a rate increase, more sellable area, or a shift into premium positions and larger packages. All three are legitimate and all three should be named, because otherwise the figure gets read as loyalty when it was pricing.

Contraction is the half nobody instruments

Of the four numbers in the bridge, three get attention and contraction does not. Churn has a call list. Expansion has a commission attached. Contraction has neither, because the account renewed and the renewal report marks it green.

Our 93 shrinking accounts cut 0.31 million between them, an average of 3,333 each, from an average base of 12,043. That is a 27.7 per cent cut on themselves. Nothing in logo retention shows it, because all 93 came back. Nothing in gross revenue retention shows it either, because gross revenue retention uses base year revenue on both sides by construction.

An exhibitor who halves a stand has usually made a decision that took months and involved somebody outside the relationship you have. Treating that as a renewal because a contract was signed is how a show ends up surprised two editions later. The practical fix is to compute the per account change and set a threshold that triggers a conversation, so that a cut of more than a quarter of previous space gets the same handling as a lapse. Which signals arrive early enough to see it coming is a separate question.

Two numbers finance can plan against

Report gross and net revenue retention on the same cohort with the four bridge components underneath, and the pair supports two different decisions.

Gross revenue retention is the floor. It is what remains if every returning account holds still, and it is the number to use when you want the downside on next edition's exhibit revenue without arguing about growth assumptions.

Net revenue retention is the actual outcome for the existing book. It is what the finance director asked for, and it should sit next to the ceiling implied by your new business plan rather than next to a benchmark from another industry.

The habit worth building is publishing the four components every edition, in the same order, with the same revenue convention on both sides of the fraction. Space only or space plus sponsorship, contracted or collected, gross rate card or net of discount: pick one and record which. The measure is a ratio of two sums, so a convention that changes between editions produces a movement that never happened.

Where this stops

The measure is contaminated by price and it cannot tell you so.

Raise the rate card 6 per cent and a returning exhibitor who takes an identical stand contributes 106 per cent of its base year revenue. On our show, a 6 per cent increase applied across the book would push net revenue retention from 90.0 to roughly 95.4 with not one extra square metre sold. Anyone reading the number as exhibitor behaviour has been misled by the pricing committee.

There are two repairs. Restate edition N revenue at edition N plus 1 prices, so both sides use one rate card and the residual movement is genuine size change. Or measure the same cohort in square metres, which removes price entirely and answers the floorplan team's version of the question. I would do the second first, because it needs no assumptions, then use the price restated revenue version for the board.

Concentration is the other limit, and it bites hardest on smaller shows. In our 512 account book the top 20 accounts hold about 1.9 million, which is 27.9 per cent of the revenue. A single anchor at 190,000 leaving moves net revenue retention by 2.8 points on its own, which is larger than most of the operational effects anyone will discuss in the review. Publish the measure with and without the top 20 when the cohort is this small, and treat a one edition movement of under three points as noise.

Multi edition contracts distort this in a way that flatters you, because an account locked for three editions cannot churn or shrink until the term ends, and that deserves its own treatment.

Take your last two editions this week and produce four totals from one join: revenue at edition N from accounts present at both, revenue at edition N from accounts absent at edition N plus 1, and within the returning set the sum of the increases and the sum of the decreases. Those four numbers are the entire bridge, and the third and fourth are almost certainly figures nobody at your show has seen written down.