Revenue concentration analysis on a show you are about to buy

The growth chart in the pack goes up and to the right. Exhibitor revenue from 7.4 million pounds three editions ago to 9.2 million at the last edition, which is 24.3 per cent over three years, and the narrative says the sector is expanding and the show is taking share.

A revenue concentration analysis on the same file takes an afternoon and asks a different question: which companies produced that 1.8 million pounds of growth. On this show, ten of them produced 1.62 million of it. That changes what the growth is worth, because growth held by ten accounts is a set of relationships, and every one of those relationships can end with a phone call.

The two numbers, and why one of them is not enough

Top ten share is the number everyone computes. On the show above, the top ten exhibiting companies held 3.77 million pounds of the 9.2 million, which is 41.0 per cent.

The problem with top ten share is that it treats the top ten as a block. Ten exhibitors at 4.1 per cent each gives you 41 per cent. So does one exhibitor at 20 per cent with nine at 2.3. Those are wildly different assets and the headline cannot tell them apart.

A Herfindahl index fixes that by squaring each company's percentage share before summing, so a 20 per cent account contributes 400 points while a 2 per cent account contributes 4. It punishes lumpiness, which is exactly what a buyer wants a concentration measure to do.

Computing the top ten share and the Herfindahl index

Aggregate booked revenue to parent company level first. A group with three brand stands is one counterparty for concentration purposes, whatever the contract table says, because one procurement decision ends all three.

On the example show, the top ten shares came out at 8.2, 6.1, 5.0, 4.3, 3.6, 3.1, 2.8, 2.7, 2.6 and 2.6 per cent. Square each and sum: 67.2 plus 37.2 plus 25.0 plus 18.5 plus 13.0 plus 9.6 plus 7.8 plus 7.3 plus 6.8 plus 6.8, which is 199.2 index points from ten companies. The remaining 405 exhibiting companies hold 59 per cent between them and contribute about 8.6 points. Total Herfindahl index, 208.

Now the same arithmetic three editions earlier. Top ten share was 29.0 per cent, the individual shares ran from 5.1 down to 1.9, and the index came to about 106.

Concentration roughly doubled while revenue grew 24 per cent. Neither figure alone would have told you that. Together they do, and the whole exercise is two pivot tables.

Where did the growth actually come from?

This is the line I would put in the report first.

Three editions ago the top ten held 29.0 per cent of 7.4 million pounds, which is 2.15 million. At the last edition they held 41.0 per cent of 9.2 million, which is 3.77 million. The difference is 1.62 million, against total revenue growth of 1.80 million. Ten companies produced 90 per cent of three years of growth.

The rest of the floor, 405 companies, grew from 5.25 million to 5.43 million. That is 3.4 per cent over three years, or barely over 1 per cent a year, which is below inflation in every market I can think of.

So the show has two businesses inside it. A small number of expanding anchors, and a broad base that is flat to declining in real terms. The seller is pricing one blended growth rate. The blend is the least informative number available.

Listed organisers make the same distinction in their own reporting for exactly this reason. Emerald Holding's full year 2025 results, published in March 2026, showed total revenues of 463.4 million dollars, up 16.2 per cent, alongside Organic Revenues of 397.0 million, up 1.1 per cent on 392.6 million. Two true numbers about the same year, answering different questions, and the gap between them is the whole point of publishing both.

Pricing the loss of one account

Concentration only matters if you can put a number on it, so put one on it.

Suppose the show does 9.2 million pounds of exhibitor revenue and 2.1 million of sponsorship, conference and other income, for 11.3 million total, and 3.6 million of EBITDA at a 31.9 per cent margin. The largest exhibitor holds 8.2 per cent of exhibitor revenue, which is 754,000 pounds.

Space revenue carries a high incremental margin, because the hall is already hired and the marketing is already spent. At 85 per cent incremental margin, losing that one account removes about 641,000 pounds of EBITDA, which is 17.8 per cent of the total. At an eleven times multiple, that single relationship is carrying roughly 7.05 million pounds of enterprise value.

Whether that is acceptable depends on the account. Ask three things about each of the top five. How long have they exhibited consecutively. Is your show their largest stand in the category, which is the exhibition brand moat question applied to one company. And is there a signed contract covering the next edition, or a handshake.

Then build the same calculation as a short table and put it in the report. One row per top five account, with revenue, share, EBITDA at risk, and enterprise value at risk at your working multiple. On this show the top five hold 27.2 per cent of exhibitor revenue, which is 2.50 million pounds. At the same incremental margin that is 2.13 million pounds of EBITDA, which is 59 per cent of the total, sitting with five counterparties. Nobody argues with a growth rate once that row is on the page, because the argument moves to which of the five is likely to leave and what the escrow should be.

A second cut is worth ten minutes. Recompute the growth story with the top ten removed entirely. Excluding them, revenue went from 5.25 million to 5.43 million and the show grew 3.4 per cent over three years. That is the number to put next to the 24.3 per cent on the front page, because between them they describe an asset the blended figure cannot.

Borrowing a regulator's threshold, carefully

The Herfindahl index has a formal life in competition law, and it is worth knowing the numbers even though they do not transfer cleanly.

The US Department of Justice and Federal Trade Commission 2023 Merger Guidelines, issued in December 2023, treat a market as highly concentrated above a post-merger index of 1,800, with a structural presumption where the merger increases the index by more than 100. Those figures describe seller concentration in a defined antitrust market, computed on market shares.

Our 208 is computed on customer shares of one show's revenue, which is a different thing entirely, and quoting the 1,800 threshold as though a show at 208 were safe would be nonsense. What transfers is the shape of the measure and the discipline of watching the change rather than the level. A move from 106 to 208 is a change of 102 points, and if you were applying the regulator's own sensitivity to changes, that is the size of move that gets attention.

Use the index as a comparator across the shows in a portfolio and across editions of one show. Do not use it as an absolute threshold, and say so in the report before somebody else does.

Why does the seller already know this?

Because they have watched those ten accounts get bigger for three years, and the sales team's forecast meeting is entirely about them.

Cuypers, Cuypers and Martin (Strategic Management Journal, 2017) studied what happens when the two sides of a deal have different amounts of acquisition experience, and found that the party with more of it captures more of the value, with the size of that effect depending on how much information asymmetry the acquirer faces about the target. An event business is an unusually asymmetric asset. The customer relationships are personal, mostly undocumented, and legible only to the people who hold them.

Which is the argument for computing concentration from raw booking files instead of asking for it. A concentration slide prepared by the seller will use a definition, an entity level and a base period, and you will not know which. Two pivot tables on the exhibitor file give you an answer nobody chose on your behalf, and they also feed straight into the quality of earnings for events work where the top accounts get tested for one off revenue and barter.

Where this stops

Concentration measures dependence. It does not measure fragility, and the two come apart in both directions.

A show where the top ten are ten global manufacturers who have exhibited for twenty years each, whose competitors all exhibit too, and whose product launch calendar is built around the event, is concentrated and extremely stable. A show where the top ten are ten regional distributors, four of whom are being acquired, is less concentrated on paper and considerably more fragile. The index cannot see any of that, and anybody who reports the number without reading the ten names is doing arithmetic instead of diligence.

The second limit is that revenue concentration ignores the audience side. A show can have a beautifully spread exhibitor base and depend entirely on two hundred buyers from six customer organisations whose attendance is a habit rather than a commitment. That dependence never appears in a booking file, and it is not something a data room will surface unless you ask about buyer organisations by name and count them the same way. The retention picture from exhibitor retention diligence is the closest available check, since a stable top ten with collapsing logo retention underneath it is the pattern that should worry you most.

Start with one pivot table. Take the last edition's exhibitor revenue, group it by parent company, sort descending, and read off the top ten share. If it is above 35 per cent, do the same thing for the edition three years earlier before your next synergy and integration review, because the direction of travel is worth more than the level.