The Association of Independent Festivals reached the twentieth UK festival of 2026 to announce a cancellation, a postponement or a fallow year on 10 June, when the Glasgow edition of WOMAD was called off over low ticket sales a few weeks before its date in the first week of July. The same trade body counted 43 for the whole of 2025 and 78 for the whole of 2024, which is the worst year it has on record, against 36 in 2023. It published the figure with three requests to government: extend the reduced VAT scheme into the 2027 season, include independent festivals in the Music Growth Package, and pilot a Grassroots Music Festival Tax Relief. Most of the coverage printed the annual numbers in a row and read a sector still falling. Those numbers cannot carry that reading, and they cannot carry the cheerful one either.
A part year and a full year are not the same measurement
Seventy-eight, 43 and 20 look like a trend because they are set beside each other. Two of them count a completed year. The third counts five months and ten days. AIF keeps its tally as a running total and revises it upward as a season goes on, which is a reasonable way to run a campaign and the worst available input to a year-on-year comparison.
The 2024 season shows how large the gap is. On 20 August 2024 the count stood at 60, reported that week by Music Week, Pollstar, IQ and AIF itself. By November the same tally read 72. The year finished at 78. A reading taken in late August 2024 therefore held about three quarters of what the year eventually produced, and a reading taken in June would have held less than that.
Twenty by 10 June 2026 says very little about how 2026 will end. It is not evidence of improvement. A partial count is being asked to do work it cannot do, and every outlet that placed it next to two completed years made the same error.
Three different outcomes are being added into one figure
The count adds together festivals that cancelled an edition, festivals that postponed one and festivals that took a fallow year. Those are not the same event. A cancelled edition may return next season. A fallow year is a planned pause, and AIF runs a support package for festivals taking one, which places it among the outcomes the organisation treats as survivable. Only some of the twenty are gone for good, and the tally does not say which.
The outlets carrying the number do not agree on what it counts. Music Ally headlined the June figure as the twentieth festival closure of the year. Pollstar wrote it as canceled, postponed or closed. Festival Insights called the same figure cancellations. Music Week used two different formulas on the 2024 count within a few months, describing 60 festivals as postponed, cancelled or closed down in August and 72 as postponed or cancelled in November. One number, four descriptions, and the difference between them is the difference between a pause and a closure.
The denominator that nobody prints
A festival can fall once. That single property is what makes a cancellation count unreadable on its own, because the pool of events available to cancel shrinks every time the count goes up. Ninety-six UK festivals were lost during the pandemic. Another 36 fell in 2023, 78 in 2024 and 43 in 2025. Whatever remained by January 2026 was a smaller and, on average, better capitalised set than the one that entered 2024.
A falling number of cancellations is what a recovering sector produces. It is also what a sector produces after the most fragile part of it has already gone. The two are indistinguishable without knowing how many festivals were operating in each year, and no such figure appeared in any of the coverage we checked, or alongside the tally itself. The count measures announcements. It was never built to measure the size of the sector, and it is now being read as though it were.
Our own index cannot see a cancellation either
We index event listings across eight cities, and the same blind spot sits in our data. Our published capture of 17 August 2026 covers 1,990 scheduled events in London across 646 distinct venues, in a 46-day window running to 1 October. Every one of those is an event that had been announced. None of them is an event that was called off.
That limit is written into the dataset we publish, which records that listings change and get cancelled after capture, and that the window is a forward view of what is scheduled rather than a record of what happened. When a promoter cancels, the listing leaves our coverage without leaving a mark, so we could not produce a cancellation rate for a city even though we hold the calendar it would be drawn from. We have run into the same wall before, when a free party boom turned out to be unmeasurable for want of a baseline. The festival number has the shape of a measurement and the substance of a press release, and so, in this respect, does ours.
What would actually settle the question
Two figures would resolve it. The first is how many independent festivals were scheduled in the UK in each year since 2019. The second is how many of those took place. A cancellation count divided by a scheduled count is a rate, and a rate can be compared across years. A raw tally of announcements cannot, whichever direction it moves.
Nothing stops that being published. Festival licensing runs through local authorities, ticketing platforms hold the scheduled set, and AIF knows its own membership. Until somebody assembles it, the accurate statement about 2026 is narrow: twenty independent UK festivals had announced a cancellation, a postponement or a fallow year by 10 June, the year was not over, and the direction of the sector was not in that sentence. The policy asks attached to the figure may well be sound. They are being argued from a number that cannot support an argument about trend, which is a weaker position than the sector deserves, and it is the sector's own count doing it.