The headline nobody wants to write

Racing likes to call itself Britain’s second most popular spectator sport. Its average crowd would finish second from bottom in League One, a division whose clubs have never claimed to be anybody’s second favourite anything.

Which brings us to 2026, and the headline I would write if anyone asked: the going’s not so good, at least so far in 2026. It lacks the swing of a fanfare, but it matches the figures.

Mind the verbs. In 2025 the Racecourse Association had attendances “top” five million. In 2026 the first half is “announced” and the third quarter “confirmed”. Trumpets are put away when the tune changes.

The first quarter rose. Then the summer arrived, and the second £3.5 million of Levy money was meant to be lifting it. The third quarter fell, the first nine months stand about two per cent down on 2025, and to scrape past five million the final quarter must beat last year’s by more than six per cent. In summer 2022 the previous campaign blamed the cost of living. This time the alibis are a heatwave, a drought and the World Cup. Racing has a gift for meeting bad luck exactly where its marketing is aimed, and at this rate the Met Office will need a seat on the GBR board.

So, three questions the headlines have stopped answering. Is the programme a good investment? Who is measuring customer retention? And how does racing invest for growth?

The return

I have spent a career in strategic consulting, so I know the slide for this moment. The client asks what they got for the money, and the next slide is called “next steps”.

The Levy Board gave Great British Racing £3.6 million for the first summer of The Going Is Good and £3.5 million for the second, about £7.2 million in all. It is a fine slogan, a double meaning on ground conditions that I wish I had thought of. I looked where a return is usually kept and found only the slogan, so here is the sum GBR has not published. Attendance in 2025 rose by about 232,000. Take out some 31,000 children, who generally go in free, and 12 per cent of adults for free tickets (my assumption), and roughly 177,000 paying adults remain.

How many came because of the campaign?

Assumes £3,620,000 spent; £30 staked per visit at a 10 per cent bookmaker margin, so 30p of levy; all of it new betting, which is generous. Attribution is my judgement, not a measurement.

My central view is 35 per cent: £58.42 to acquire each extra paying adult, and 30p a visit coming back. He must go racing 195 times to repay it, nearly four years of Saturdays, a loyalty scheme no airline would dare.

That is why retention is the whole game. £58.42 is a good investment only if he comes back, and racing counts visits, not visitors: a man who comes twelve times is twelve heads. It counts heads the way a shepherd counts sheep, which is fine until he wants to know which came back. A vendor’s 2017 case study for the RCA and GBR reported that 73 per cent of purchasers did not buy year on year. Treat it gently, but it shows the question can be asked. I found no retention figure in any RCA release.

Who pays, and who collects

Now the consultant’s question: who funds the return, and who collects it? The Levy Board buys the advert with a 10 per cent levy on bookmakers’ gross win, about £110 million a year. The racecourse sells the ticket and keeps the gate and the bar. The Levy Board’s share of the new customer is the 30p. It is the only supply chain I know in which the man paying for the shop window never sees the till, and the man who owns the till never saw the advert.

The insight the headlines miss is that nobody owns retention. The funder cannot bank the gate, and the courses, who can, compete and keep their data apart. In the vendor study only 6 per cent of buyers in 2014 bought at more than one course, so the customer is loyal to a venue, not to racing. A national campaign for racing is rather like advertising “pubs” to a man who drinks only in the Red Lion. The Red Lion is delighted, and the man who paid for the advert never learns whether he came back.

Then there is Goodhart’s law, the economist’s way of saying that a measure which becomes a target stops measuring. The five-million milestone is the target, the turnstile is the measure, and a free ticket moves it as well as a paying one. The publican calls this a free pint behind the dominoes league.

To be fair, the Levy Board is not poor: its yield has risen five years running. But betting turnover per race is down nearly a fifth since 2021/22, so it is buying footfall in a shrinking betting base and cannot show that one has anything to do with the other.

Saying the sentence

Change works a bit like it does with the alcoholic: nothing happens until you stand up and say the sentence. Racing is short of denominators, not explanations. The sentence is not “the weather was dreadful” but “we do not know what our crowds are, or what they are worth”. I have owned racehorses, so I recognise the conversation: it is the one with the trainer about why the ground was against us. I say this as someone who has studied form for fifty years and would like to go on enjoying it.

What is actually counted

I do not dispute 2025. It was the first year above five million since 2019, and the average crowd rose. But 2019 was 5.62 million, so even the good year was about 10 per cent short, and the first half of 2026 is 14 per cent short. Courses now count owners by attendance, not badges issued, so the old number was generous.

And what is in the number? About 31,000 of the extra heads were children. Nobody publishes the free adults: owners’ guests, sponsors’ guests, ballot winners, walk-ins who never meet a till, the couple given a pair by the local radio station. Add the weather, 26 fewer abandonments, worth roughly two-fifths of the increase on my arithmetic. Dry Saturdays with a marketing budget attached.

Back to that second place. It is a claim about total admissions, not spectators per fixture, and per fixture the average of 3,526 sits between Burton Albion and Stevenage. Wolverhampton managed 629 a fixture in 2023, and some fixtures, paying and non-paying together, have barely made a hundred. That is not a crowd. That is a committee meeting with a bar. The racecourse clicker counts the owner’s guest, the estimated member, the free child and the paying adult, and gives each one. Nothing is wrong with the clicker. The trouble is the question.

Project Beacon: a lighthouse with no harbour

Then there is Project Beacon, backed by £484,000 from the Levy Board. I called it a lighthouse pointing at the obvious and read the follow-up hoping to be wrong. M&C Saatchi’s work is beautifully made. It sizes the market at 25 million adults and divides them into eight segments with names that would grace a Sunday supplement, Hesitant Onlookers and Side-line Socialisers among them, then sets six “jobs to be done”. Only a consultancy could make a to-do list sound like a strategy. The list has six items and no dates.

The Beacon summary states no quantitative targets, KPIs or success measures. It marks three segments for retention, Serious Betters, Event Lifers and Racing Experts, and the BHA’s first-quarter update reports no retention rate for any of them. It reports 100 million video views at Cheltenham, not how many viewers became visitors. That update treats any link to the first-quarter rise as early indications, written when the only quarter available was the good one.

A target never set can never be missed. That is the genius of it: a strategy incapable of failing, which in a press release looks identical to succeeding. A lighthouse exists to stop ships hitting the rocks. This one has produced a handsome survey of the rocks.

How racing should invest for growth

From the documents, racing invests by buying attention: advertising, research, and course marketing on top. It buys the first visit and has not said what it spends on the second. Three next steps, which is the traditional number, are my judgement and untested.

Count the customer, not the click. One customer reference across all 59 courses, so the first-timer of 2025 can be found in 2026. Publish paid admissions and the share who return.

Buy the second visit. Offer first-time buyers a return ticket, valid at any course within 60 days, tested against a control group and priced well below £58.42. Ayr’s racing club and Newbury’s free horse share, both in Beacon’s update, are the right instinct, but I found no published result.

Make the collector pay. Match-fund the national campaign from the courses in proportion to measured paid-admission gains, and release the Levy Board’s share against retention targets, so whoever owns the till also owns the advert.

Where the messaging runs ahead of the facts

The baseline that moves. Q1 2025 was down on 2024, so the headline compared it with 2023, because Easter moved. Fair. Yet 2026’s nine months are down on 2025 and up on 2024, and the upward comparison gets the airtime.

The credit nobody audited. GBR’s chief customer officer said no gain could be attributed exclusively to the campaign, and the BHA’s chief executive called it “a contributing factor”. Website traffic was reported up 60 per cent; traffic is not tickets, and no return per pound was claimed.

In fairness, take out every child and 2025 was still a few per cent up, which I have checked twice and dislike. Racing has more customers. It cannot say how many it bought.

The going, I am told, is good. I should like to see the receipt, and who came back.

Sources

• RCA: racecourse annual attendances top 5 million (2025 statement)

• RCA: Q1 2025 attendance returns

• RCA: first half of 2025

• RCA: third quarter 2025

• RCA: half year 2026 attendances confirmed

Figures marked as my arithmetic or my assumption are derived from the sources above, not published by them.