The largest direct cost lines on a trade show and how they move

The budget review is forty minutes in and somebody asks what the largest direct cost lines actually are. The answer that comes back is a 400 row export from the finance system, sorted alphabetically, with account codes instead of names. Nobody in the room can tell from it whether the show is expensive because of the hall, the build or the labour.

That export is the problem. A show has four cost lines that decide the answer and roughly forty that do not, and the four are usually buried among the forty because the chart of accounts was designed for the group rather than for a show director.

What counts as a direct cost line?

Direct cost is the cost that would disappear if the show did not happen. Hall hire disappears. The show director's salary does not, at least not this year, and neither does the group marketing team or the office.

Emerald Holding sets this boundary explicitly in its Form 10-Q for the quarter ended 30 June 2025, where it says that direct trade show costs are recorded in cost of revenues and all other costs are recorded in selling, general and administrative expenses. The same filing names the categories inside cost of revenues: decorating expenses paid to general service contractors, sponsorship costs owed to endorsing trade associations, venue costs, costs of other marketing services, and other event related expenses covering temporary labour for security, shuttle buses, speaker fees, food and beverage, and event cancellation insurance.

That list is worth reading twice, because it is a listed organiser telling you in a filed document what its own show cost stack looks like. Notice what is in it and what is not. The people who sell the space are in selling, general and administrative expenses. The people who build the aisles are in cost of revenues.

The four lines that carry the weight

Take a mid sized business to business show with 3.1 million of direct cost. A stack that looks like this is ordinary.

Floor build and decorating, 810,000. That is the general service contractor: aisle carpet, entrance features, hanging signage, rigging, communal area furniture, the graphics package. Hall hire and venue services, 640,000, which is the space itself plus the utilities, the in house rigging fees and whatever the venue insists on selling you. Contracted operations labour, 520,000, covering security, cleaning, shuttle buses and the temporary staff on the registration desks. Audiovisual and production, 390,000, for the keynote stage, the session rooms and the technicians who run them.

Those four sum to 2,360,000. Against a 3.1 million base that is 76 per cent of direct cost. The remaining 740,000 is spread across roughly forty lines, which averages 18,500 each.

The arithmetic matters more than the specific figures. Any hour you spend on the 740,000 is an hour spent on lines averaging under twenty thousand, where a heroic 10 per cent saving on one of them recovers 1,850. The same hour spent on the build line recovers 81,000 at the same 10 per cent. Cost work on a show is almost entirely a question of where you point it.

Why does the venue line vary so much between organisers?

Here is where borrowed benchmarks fall apart. In the same 10-Q, Emerald Holding writes that convention centres are typically owned by local governments who have a vested interest in stimulating business activity in and attracting tourism to their cities, and that venue costs therefore typically represent a small percentage of its total cost of revenues.

For a US organiser running in city owned convention centres competing for room nights, that is credible and it makes hall hire a minor line. For an organiser running in a privately operated European hall, or in a market where a single venue has no competition within four hundred miles, hall hire can be the largest single line on the show. The cost structure follows the venue market, and the venue market is local.

So the honest version of the ranking is that floor build and hall hire trade places depending on where you run, and the two labour driven lines sit underneath them in most portfolios. If you have inherited a cost model from a group that runs somewhere else, check that assumption before you plan against it. The venue clauses that move the hall line after you have signed are their own subject and not this post's.

How each of the four lines moves

Now the useful part. Suppose the floor grows 12 per cent, from 31,000 net square feet sold to 34,720.

Floor build moves close to proportionally, because carpet, signage and rigging are bought by the metre and the square foot. Take 810,000 up by 12 per cent and you get 907,200, an increase of 97,200. Not exactly proportional, since the entrance feature and the registration build are fixed whatever the floor does, but proportional enough to plan with.

Hall hire moves in steps. If the extra 3,720 square feet fits inside the halls you have already booked, the line does not move at all. If it forces a third hall, the line jumps by whatever that hall costs, perhaps 180,000, in a single move that no percentage assumption would have predicted. This is the line that breaks naive budget models, because the finance team applies a growth rate to it and the venue applies a price list.

Contracted operations labour moves partly with area and partly with opening hours. More aisles need more security posts and more cleaning passes, so part of the line follows the floor. The bigger driver is often the hours: extending the show from two days to three adds a full day of every post, which can move the line more than a 12 per cent floor increase does.

Audiovisual is driven by the session count and the stage count. A show can add 12 per cent of floor and not touch the audiovisual line at all, and then add one extra theatre and put 60,000 on it. Treating audiovisual as a variable cost against floor area is one of the more common modelling errors in event budgets, and it produces forecasts that are wrong in both directions.

The point for planning is that only one of the four largest lines behaves the way a spreadsheet assumes. Splitting these properly into fixed and variable behaviour is a modelling job of its own.

What do you do with the forty small lines?

Mostly, leave them alone and control them by rule instead of by review.

The 740,000 tail on our example show contains the badge stock, the lanyards, the water for the aisles, the first aid cover, the floral, the golf carts, the crew catering, the radio hire, the waste removal surcharges and thirty more like them. Individually they are too small to justify a negotiation. Collectively they are 24 per cent of direct cost, and in a bad year they are where the overrun hides, because each one is small enough that approving it feels harmless.

The control that works is a threshold with a name attached. Every line under, say, 25,000 gets a fixed budget set from last edition plus an inflation assumption, and any request to exceed it goes to one named person rather than to whoever is standing nearby on build day. That converts forty negotiations into one policy. It also produces a clean variance report, because a line that was set by rule and then broken tells you something, while a line that was never set tells you nothing.

The exception is the small line that is growing fast. Waste removal and utilities have moved sharply at several venues over recent contract cycles, and a line that was 14,000 two editions ago and is 31,000 now has stopped being a tail line even though it still looks like one in an alphabetical export. Sort the tail by year on year change, not by size, and the two or three worth attention surface immediately.

What margin band are these lines defending?

Cost work feels abstract until you see how thin the profit band around it is. UFI's Global Exhibition Barometer, in the 37th edition published in July 2026 from a survey of 466 companies in 59 countries, found 19 per cent of companies expecting operating profit to increase by more than 10 per cent in 2026, while 54 per cent expected profit to stay within plus or minus 10 per cent.

Read that as a distribution rather than a forecast. Most organisers expect this year to look roughly like last year at the profit line. On a show carrying 3.1 million of direct cost, a 6 per cent overrun on the build line alone is 48,600, and two of those in the same edition will move a show from the stable majority into the declining tail without anyone making an obviously bad decision.

Where this stops

Ranking cost lines tells you where to look. It does not tell you what is avoidable, and the gap between those two is where most cost programmes die.

The build line is the largest and the least compressible in the short term, because the general service contractor contract is usually multi year and the specification was set by the show's own brand standards. The hall line is large and effectively fixed once the contract is signed, which is often three to five years out. The labour line is genuinely negotiable, but it is also the line where cutting produces visible degradation on the floor, and an exhibitor complaining about aisle cleaning on day two costs more in rebooking than the saving was worth.

There is also a coverage limit. This ranking assumes your finance system tags cost to a show at all. Plenty of portfolios book shared costs such as a portfolio wide audiovisual contract to a central code and allocate later, which makes any single show's cost stack partly an artefact of the allocation rule. If that is your situation, the ranking above is still directionally right and the numbers are not yours yet.

Start by exporting last edition's direct cost to a single sheet and forcing every row into one of five buckets: build, venue, operations labour, audiovisual, and everything else. Do not create a sixth bucket. Then check what share the top four carry and whether the fifth bucket is small enough to ignore, which will tell you whether your cost model is worth arguing about or whether the chart of accounts needs fixing first. Once you have that split, the event finance work that follows it, starting with contribution per square foot and the gross margin the board reads, has something real underneath it.