The biennial event cycling adjustment that makes odd and even years comparable

Somebody plots five years of portfolio revenue and the chart has teeth. Up eleven, down nine, up twelve, down eight. The commercial director asks what happened in the down years and the honest answer is that nothing happened, which is a difficult thing to say out loud in a portfolio review.

What the chart is showing is the calendar. A biennial event cycling adjustment is the correction that takes the calendar back out, and it is worth building carefully, because the two obvious ways of doing it answer different questions and one of them throws away a show.

What is the biennial event cycling adjustment doing?

Making both sides of a growth comparison contain the same set of scheduled editions.

Informa PLC states its version in the glossary of its 2025 full year results: "Phasing and biennial adjustments relate to the alignment of comparative period amounts to the usual scheduling cycle of events in the current year." The wording is worth reading twice, because it says the comparative moves. The current year is taken as given, and the prior period is restated to hold the same shape.

RELX PLC does the same thing at the level of its definition, stating in its 2025 annual report on Form 20-F that underlying revenue growth rates "are calculated at constant currency and exclude revenue from acquisitions until twelve months after purchase, revenue from disposals and assets held for sale, print and print-related revenue and exhibition cycling". Cycling is one of four named exclusions, sitting alongside acquisitions and currency, which tells you how large the sector considers it.

Both groups are solving the same problem. A portfolio containing shows on a two year cycle reports growth in the years those shows run and a decline in the years they do not, and neither number describes how anybody traded.

The arithmetic on a portfolio with one biennial

One biennial in a portfolio of a dozen annual shows is enough to break the series. Here is the size of it.

The annual shows produced 35.1 million in the odd year. In the even year they produced 34.4 million, which is a decline of 0.7 million, or 2.0 per cent. The biennial ran in the even year and brought in 4.6 million.

Reported revenue for the even year is 34.4 plus 4.6, which is 39.0 million. Against the 35.1 million comparative that is growth of 3.9 million, or 11.1 per cent.

So the pack says plus 11.1 per cent and the trading says minus 2.0 per cent, a gap of 13.1 percentage points created by one show that was always going to run.

Now roll forward a year. The biennial does not run. The annual shows have a decent year and grow 3 per cent, from 34.4 million to 35.4 million. Reported revenue is 35.4 million against 39.0 million, a decline of 9.2 per cent.

Two consecutive reported figures, plus 11.1 per cent and minus 9.2 per cent, a swing of 20.3 points, describing a business whose annual shows went from minus 2.0 per cent to plus 3.0 per cent. The reported series is four times as volatile as the thing it claims to measure, and the volatility is entirely predictable in advance, which is what makes leaving it in indefensible.

Two ways to build the adjustment, and which one to use

The first method drops the biennial from both periods. Underlying growth then covers the annual shows only, and the arithmetic is the 34.4 against 35.1 above.

The second method restates the comparative to the current year's cycle, which is what the Informa wording describes. The biennial ran this year at 4.6 million, so its previous edition, two years ago, is pulled into the comparative at whatever it produced. Say that was 4.2 million. The comparative becomes 35.1 plus 4.2, which is 39.3 million, and the current year stays at 39.0 million. Underlying growth is minus 0.8 per cent.

Those are different answers to different questions, and I would use the second one.

Dropping the biennial answers how the annual shows performed. That is useful and incomplete, because a biennial can easily be a tenth of portfolio revenue, and a measure that reports nothing about a tenth of the business in the year it actually ran is a strange measure to put in front of a board.

Restating the comparative keeps the biennial in and measures its current edition against its own last edition. In the numbers above that edition grew from 4.2 million to 4.6 million, which is 9.5 per cent over two years. Compounded, that is about 4.6 per cent a year, which is below what the annual shows managed in the good year and above what they managed in the bad one. The show is in the conversation, which is where it should be.

The cost of the second method is that the comparative is two years old and carries two years of rate card increases, inflation and market movement inside a single growth rate. Say so when you publish it. A biennial's growth rate is a two year number wearing a one year label, and a reader who does not know that will compare it directly against the annual shows and draw the wrong conclusion.

Why does cycling move profit more than revenue?

Because a biennial edition arrives with its direct costs and almost none of the central ones, so an unusually high share of its revenue reaches the bottom line.

The published evidence for this is in the same Informa statement. The reconciliation from underlying to reported growth carries a phasing and other items column for both revenue and adjusted operating profit. In 2025 that column was 1.3 per cent on revenue and 2.4 per cent on adjusted operating profit. In 2024 it was negative 3.4 per cent on revenue and negative 7.7 per cent on adjusted operating profit.

In both years the profit effect is roughly double the revenue effect, and in 2024 the gap is 4.3 percentage points of profit growth. The sign flips between the two years as well, which is the sawtooth showing up in a published document rather than in a chart somebody made.

The practical consequence for an organiser is that the profit forecast for a biennial year should carry a higher drop through rate than the portfolio average, and the year after should carry a fall in profit that is proportionally larger than the fall in revenue. Teams that model the revenue cycle and then apply an average margin to it will under-forecast profit in the on year and over-forecast it in the off year, twice, every cycle.

Getting the cycle out of people's heads and into the events table

The adjustment fails in practice for a boring reason, which is that nobody records the cycle as data.

Put a cycle field on the event record with the frequency in months, alongside the date of the previous edition and the expected date of the next. Twelve for an annual, twenty four for a biennial, thirty six for a triennial. Then the comparative restatement is derived rather than remembered, and it survives the analyst who built the spreadsheet moving teams.

Two cases break a naive version of that field and are worth handling explicitly. A show that changes cycle, usually a biennial going annual because the market got busy enough to support it, has a transition year where the previous edition is twelve months back rather than twenty four. And a show that rotates between cities on a two year cycle is a biennial in each city and an annual at portfolio level, so the adjustment depends on whether you are reporting the brand or the edition.

Neither is hard. Both are impossible if the frequency lives in somebody's memory. The set of editions that ends up in the comparison is the same object that a bottom up like for like comparison is built from, so the field earns its keep twice.

Where this stops

The adjustment assumes the biennial's previous edition is a fair comparative, and across a two year gap that assumption gets weak.

A biennial whose last edition ran in a different economic climate, at a different venue, or before a competitor launched against it, is being compared against a number that describes a different market. The adjustment cleans up the calendar effect and it cannot clean up the two years of change sitting inside the comparative. Where those two years contained something specific and large, the honest treatment is to publish the underlying growth rate with a note naming what changed, rather than presenting it as though it were an annual comparison.

There is also a boundary worth keeping firm. A cycling adjustment covers shows that were always going to run on that pattern. An edition that moved because a venue was unavailable or a date clashed is a different problem with a different fix, and an unplanned move between financial years can put two editions in one year rather than none. Informa is explicit in the same glossary that its adjustment does not extend to editions cancelled or postponed, which were left to hit underlying growth unadjusted in 2024 and 2025. Cycling and disruption look similar in a chart and should be treated separately in a definition, which is one of the reasons a group organic growth definition lists its exclusions individually.

The first step this week

List every show in the portfolio with its cycle length in months and the date of its previous edition, then mark which financial years each one lands in for the next four years. Total the revenue by year on that basis. If the resulting line has teeth in it, you now know exactly which years need the adjustment and how big it will be before anyone has to explain it in a meeting. Keep the list with the finance reporting definitions so the phasing line in next year's pack is derived from it.