Reading the registration plateau mid campaign before you spend more money

The request came in on a Thursday afternoon. Five weeks of flat registrations on a security and fire safety show, a marketing director asking for 80,000 to be moved into paid social by Monday, and a show director ready to sign it because the alternative was watching the line sit still.

A registration plateau mid campaign is where unplanned spend gets authorised, and nobody in that conversation had looked at what the same five weeks had done in the previous three editions. Somebody did, it took about twenty minutes, and the 80,000 stayed where it was.

What do people mean when they say the curve has gone flat?

The cumulative registration line never actually goes flat. It goes slow, and slow looks flat on a chart whose y axis runs from zero to twelve thousand.

Be precise about the object. A plateau is a stretch of the campaign where the weekly registration count drops to a low, roughly constant rate and stays there for several weeks. On a cumulative chart that is a shallow diagonal. On a weekly chart it is a run of similar bars. The weekly chart is the one to look at, because it turns the thing you are arguing about into a number instead of a slope you are eyeballing.

The panic is real and it has a mechanical cause. The plateau lands in the part of the calendar where the campaign is spending most heavily, so the team can see money going out and very little coming back, week after week, with no way to tell whether the money is failing or the calendar is.

The middle of a campaign is arithmetically quiet

Before you compare anything, it helps to know roughly how much of a registration file is even available to arrive in the middle.

Maritz's Registration Insights Report 2024, built from more than 360,000 attendee registration records across 30 trade shows over a three year period, put 45 per cent of attendee registrations inside the final four weeks, with the remaining 55 per cent arriving earlier than that. The final week alone accounted for 13 per cent of the file and on site registration for another nine.

Run those proportions against a show closing on 12,000 registrations with a 26 week campaign. The 55 per cent that arrives more than four weeks out is 6,600 registrations spread across 22 weeks, which is 300 a week. The final week delivers 13 per cent of 12,000, which is 1,560. The last seven days of the campaign run at more than five times the average weekly rate of the entire period before it.

It is worse than that inside the middle, because the front of the campaign is loaded. If the opening fortnight takes a fifth of that 6,600, the remaining twenty weeks deliver 5,280, which is 264 a week. A show adding 264 registrations a week against a 12,000 target is on plan and looks moribund. The quiet middle is a property of the shape of the whole campaign curve, which A2 takes on directly.

The test worth running before you authorise anything

The test is one comparison and it needs data you already have.

Take the window you are worried about, expressed in weeks to show open rather than dates. Compute the registrations added inside that window for this edition and for the same window in each of the last three editions. Then normalise, because the editions are different sizes and a raw count comparison will mislead you in whichever direction the show has been growing.

Here is the security show, weeks minus 20 to minus 12, an eight week window.

This edition held 5,120 registrations at week minus 20 and 7,360 at week minus 12. It added 2,240, which is 280 a week.

The 2025 edition went from 5,460 to 7,604, adding 2,144, which is 268 a week. The 2024 edition went from 5,010 to 7,338, adding 2,328, or 291 a week. The 2023 edition went from 4,700 to 6,900, adding 2,200, or 275 a week.

Normalise against the count standing at the start of the window. This edition grew by 2,240 on a base of 5,120, which is 43.8 per cent. The three prior editions grew by 39.3 per cent, 46.5 per cent and 46.8 per cent on their own bases. The current edition sits inside that spread, closer to the middle of it than to either end.

Do it once more against final registrations, which you know for the three closed editions. Those finals were 10,600, 11,050 and 11,400. The window delivered 20.8 per cent, 21.1 per cent and 18.8 per cent of the eventual file. Weeks minus 20 to minus 12 have contributed about a fifth of every recent edition, at a weekly rate in the high 200s, and they did it again this year.

The flat is the pattern. There is nothing in that window to fix.

What the comparison is allowed to conclude

It concludes one thing: the plateau is not evidence of a problem. That is narrower than it sounds and it is the claim people actually need on a Thursday afternoon.

It does not conclude that the campaign is healthy, that the final number will land where you want it, or that the media plan is working. A plateau sitting exactly on its historic rate is consistent with a show that is 8 per cent down overall and has been since launch, because the window comparison is a comparison of slopes and says nothing about the level the slope started from. Read the level separately, as a pacing index, which is A1's subject.

Freeman's end of year trends recap, released in January 2026 from its 2025 research, found that 50 per cent of the events it tracked were behind their regular registration pace that year. Half the industry spent 2025 looking at a number lower than the one it expected. In a year like that, the base rate for a flat middle being a genuine emergency is lower than the room believes, and the burden of proof should sit with the person asking for the money.

When is the flat actually a signal?

Three patterns turn a plateau into a signal, and all three are visible in the same weekly table.

The plateau starts earlier than it should. If your last three editions ran at 270 a week from week minus 20 and this edition dropped to that rate at week minus 26, you have lost six weeks of a higher rate at the front, and the deficit is banked. Nothing later in the campaign gives it back, because the later weeks are already running at their own historic maximum.

The rate sits below the whole prior spread rather than inside it. A window growing 31 per cent against a prior spread of 39 to 47 per cent is outside the band, and one observation outside a band of three is worth a morning of investigation before it is worth a budget.

The composition moves underneath a normal total. A window that delivers its usual 280 a week, made up of exhibitor guest passes where it used to be self-registered buyers, is a worse window with an identical shape. The total is the last thing to move and the first thing anyone looks at. Splitting the weekly count by registration type is the cheapest attendee analytics you can run against a plateau.

What I would do with the money instead

My own position is that unplanned mid-campaign spend is close to the worst use of an audience acquisition budget, and the plateau is where it always gets proposed.

The reason is timing rather than channel quality. Media authorised in week minus 18 buys registrations that would have been cheaper to buy in week minus 30 and are competing for attention against a decision the buyer has not been prompted to make yet. Maritz's 2024 figures put the heaviest registration weeks at the end of the campaign, so money spent to accelerate the middle is trying to move volume out of a period where it was going to arrive anyway.

If a genuine deficit exists, the honest options are to spend earlier next edition, to spend into the final six weeks where the file is actually forming, or to accept the number and tell operations. Each of those has to clear a cost per registration test that belongs to B39 and B40, and none of them is improved by being decided in a hurry.

UFI's Global Exhibition Barometer, 36th edition, published in January 2026 from a survey of 378 companies across 57 countries and regions concluded that December, found the state of the economy in the home market to be the most pressing short term business issue globally, named in 19 per cent of answers. When a plateau is showing up across your portfolio at the same time, the cause is usually sitting above the campaign rather than inside it, and no amount of paid social fixes a procurement freeze.

Where this stops

Three prior editions is a thin comparison and I would not pretend otherwise. Three observations give you a spread, not a distribution, and a current reading that falls just outside three prior values is a weak result. Building something you can actually put a percentile on needs five editions and a band with percentiles on it, which is A7's method.

The comparison also assumes the campaign windows are alignable. If this edition opened registration nine weeks later than the last one, week minus 20 is not the same point in the two campaigns and the whole exercise is measuring a scheduling decision. Check the registration open date in days to show open for every edition before you compute anything, and check where the price tier boundaries sit, because a tier deadline inside one edition's window and outside another's will produce a difference that has nothing to do with demand.

The deepest limit is that a plateau is a fact about timing and the question being asked is about outcome. Nothing in a window comparison tells you the show will close where you need it to. It tells you this particular flat stretch has happened before, which is enough to stop a bad decision and not enough to make a good one.

This week, build the weekly table for one show. Registrations added per week, for the current edition and the three before it, aligned on weeks to show open, with each edition's final registration count in the header. Print it on one page. The next time somebody says the campaign has stalled, you will be able to answer within the meeting rather than after it.