Logo retention versus revenue retention and why you report both

A show director I know keeps logo retention versus revenue retention on the same slide because of an argument he lost. His logo retention had been 70 per cent for three editions and his exhibit revenue had grown every year, and a board member asked him to explain how both could be true. He could not, on the spot, and it took a week to work out that the answer was flattering: the companies leaving were small and the ones staying were buying more space.

The next year the same two numbers moved the same way and the answer was the opposite. That is the case for reporting both.

What is the difference between the two measures?

Fix a cohort. Every distinct exhibiting company present at edition N, on whatever entity and date rule you have already written down, with the space revenue each of them contracted for at edition N. Then look at edition N plus 1 and mark each account present or absent.

Logo retention is the share of accounts present again. Count the heads, divide, done.

Gross revenue retention is the share of edition N revenue that belonged to accounts present again. Same cohort, same presence test, weighted by what each account was worth in the base year. It answers a different question: of the money we had, how much of it is attached to a relationship that survived.

Notice what gross revenue retention deliberately leaves out. It uses the base year's revenue on both sides, so an account that returns and doubles its stand counts exactly as much as one that returns and halves it. That makes it a clean measure of relationship survival, weighted by size, and it cannot exceed 100 per cent. Expansion is a separate measurement with its own arithmetic and its own post.

The arithmetic on one show

640 exhibiting companies at edition N, contracting 4.2 million in space revenue between them, so an average of about 6,563 per account.

448 of those companies come back at edition N plus 1. Logo retention is 448 divided by 640, which is 70.0 per cent.

Those 448 accounts had contracted 3.7 million of the 4.2 million at edition N. Gross revenue retention is 3.7 divided by 4.2, which is 88.1 per cent.

Eighteen points separate the two numbers, and the whole of the gap is size. The 192 accounts that left held 0.5 million between them, an average of 2,604 each. The 448 that stayed averaged 3.7 million over 448, or 8,259 each. Your returning exhibitors were on average 3.2 times the size of your departing ones.

That is a good result and it is worth saying so plainly, because a 70 per cent retention rate read on its own looks like a show in trouble. Read with the revenue weighting, it is a show shedding its smallest accounts while holding the ones that pay for the venue.

The gap can run the other way

Keep logo retention at exactly 70 per cent and change only who left.

Suppose the same 192 departures had held 1.4 million of the 4.2 million, an average of 7,292 each, and the 448 survivors held 2.8 million, an average of 6,250. Logo retention is still 448 over 640, 70.0 per cent. Gross revenue retention is 2.8 over 4.2, which is 66.7 per cent.

Two shows, identical headline retention, and a 21 point difference in the money that survived. The second show is losing anchors and replacing them with nothing, and there is no way to see it in a count of companies.

Concentration makes this sharper than most people expect. In our 4.2 million, the top 40 accounts hold 1.6 million, which is 38.1 per cent of the revenue from 6.3 per cent of the logos. Lose three of those 40, worth 140,000 between them, and logo retention falls by 3 over 640, which is 0.5 of a point. Gross revenue retention falls by 140,000 over 4,200,000, which is 3.3 points. The same three accounts move one measure seven times as far as the other.

Any show with a concentrated top end has this property, and most B2B exhibitions do. A count-based retention rate is systematically insensitive to exactly the accounts whose loss would end your year.

Why does value weighting change which accounts matter?

The academic version of this argument is worth knowing, because it goes further than reporting.

Ascarza, Neslin, Netzer and colleagues, writing in Customer Needs and Solutions in 2018, set out an integrated view of retention management and press repeatedly on a distinction that most retention programmes collapse: identifying which customers are at risk is a different problem from deciding which customers to act on. A programme built on risk alone will spend its effort on the accounts most likely to leave, and those are frequently the smallest, the least persuadable and the least valuable to keep.

Value weighting is the first correction to that. Once you rank by revenue at risk instead of probability of leaving, the top of your list changes composition immediately. An account at 20 per cent risk holding 84,000 of space carries 16,800 of expected loss. An account at 70 per cent risk holding 6,000 carries 4,200. The second account is three and a half times more likely to go and worth a quarter as much to save, and a risk-sorted call list puts it first.

The reporting consequence is that logo retention is the wrong number to give a sales director as a target, because it makes 192 small departures look like a bigger problem than three anchor departures. Give them both, with the revenue-weighted one governing the call list.

Reichheld and Sasser made the underlying economic case in Harvard Business Review in 1990, reporting that a 5 per cent cut in the defection rate generated 85 per cent more profit in one bank's branch system and 30 per cent more in an auto-service chain. That effect runs through margin retained, which is a revenue-weighted quantity, so the measure that best tracks the profit story is the one weighted by what the accounts were worth.

The industry's own index reports both, and they diverge

There is a useful confirmation of all this in published data.

The CEIR Index, reported through IAEE in 2025, tracks four components for the same event population, and the second quarter of 2025 showed them 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. The count of exhibitors and the real revenue those exhibitors produced moved almost 7 points apart over the same period, on the same sample.

That divergence at industry level is the same phenomenon as the gap in your own file, running the unfavourable way. Counts have held up better than money. A show measuring only logos across those years would have reported a milder decline than its finance function experienced, which is a reasonable description of a lot of post-2019 board conversations.

Reporting the pair without confusing anyone

Two numbers on one slide invites the question of which is real, so name them and give each a job.

Logo retention is the operational number. It counts relationships, it is stable, it needs no revenue join, and it can be produced on any show in the portfolio including the ones with messy financial data. It is the number to use for cohort work, for tenure curves and for comparing across shows.

Gross revenue retention is the commercial number. It is what finance should plan against, what a portfolio roll-up should sum, and what sorts the call list any renewal intelligence work produces.

Publish the gap between them as its own figure, because the gap has information in it. Eighteen points means your leavers are small. Three points means your leavers look like your average account. A negative gap means your leavers are larger than average, and that is the one to escalate the day you see it, regardless of what the headline retention rate says.

Where this stops

Both measures share a blind spot, and it is a large one. They evaluate presence, so an account that returns at half its previous size counts as fully retained in both.

On our cohort, if all 448 returning accounts came back with 15 per cent less space each, logo retention stays at 70.0 per cent, gross revenue retention stays at 88.1 per cent, and the show loses about 555,000 of revenue that neither number will show you. That is the entire reason net revenue retention exists as a separate measure, and it belongs in a separate discussion.

Gross revenue retention has a second, quieter problem: it depends on how you allocate revenue to an account, and organisers do this inconsistently. Space only, or space plus sponsorship? Gross rate card or net of discount? Contracted or collected? A sponsorship-heavy exhibitor can look twice the size under one convention and average under another, which moves the weighting and moves the rate. Pick one convention, apply it to both editions, and record it next to the definition. The convention problem disappears entirely if you weight the same cohort by square metres, which is a cleaner instrument and a different post.

Take one show's last two editions this week and compute the two numbers on the same cohort: accounts back over accounts present, and base-year revenue of accounts back over total base-year revenue. Then compute the average value of the accounts that left and the average value of the ones that stayed. The ratio between those two averages tells you in one number whether your churn is costing you money or tidying your floorplan.