Without overly dwelling on Formula 1’s recent past, it is fair to state that under previous owners CVC Capital Partners the focus on was on the short-term creaming of F1’s bottom line, and seat-of-pants management under then-CEO Bernie Ecclestone, rather than on organically growing the sport.
Marketing? Heck, that was mainly in the hands of teams, TV broadcasters and race promoters, while the media was expected to offer a free PR service.
That all changed dramatically when CVC sold up to Liberty Media. The new owners instituted a classic marketing approach, one drawing heavily on the Seven Ps of marketing: Product, Price, Promotion, Place, Packaging, Positioning and – crucially – People.
Advances in technology and the explosive growth of social media offer new opportunities to understand the latter. The days of ‘suck-it-and-see’ are over; Liberty uses a wholly scientific approach to interpreting the reactions of fans – both spectators and sofa-surfers.
Matt Roberts, F1’s global research director, presented the research methodology to a trio of media outlets including this publication in Singapore. Also present was F1 technical director Pat Symonds, who explained how the findings informed the sport’s future direction.
As an example, an ongoing bone of contention is the subject of grid penalties for drivers who use too many engine components. F1 used its Fanvoice website, which provides a base of over 80,000 fans who participate in surveys and polls, to quiz the sport’s followers on the subject.
Around 8,000 responses were received. “That is,” says Roberts, “a very robust sample size, given that political polling is only a thousand people across the country in the UK.
“We do a whole range of things. We do lots of work with our spectators – we do spectator experience research, we track people around the track, we’ve launched [customer relationship management] CRM Data Space, and we’ve launched this kind of data analytics.
“All of what we do goes not just internal[ly], we also work closely with all the teams, with the promoters, with (F1) partners and send them all the stuff we do that is relevant for them.”
Advert | Become a RaceFans supporter and
This enables F1 to understand many things about its fans: Who they are, where they are, how old they are, their gender, why they are watching, and how they value partners, drivers and teams.
“We’ve launched recently a kind of an analytics function,” adds Roberts. “Before it was more primary research surveys, et cetera, but in January 2019 we launched a data analytics team and the goal really was using big data to help drive, to help inform company decisions, and particularly around the motorsport team, the broadcast team and so forth.”
F1 also tracks conversations on various social media platforms such as Reddit and via biometrics in order to understand TV viewing habits.
“We’re diving deep into our TV ratings to understand what drives up our markets to assist our TV rights team.
“For example, we have somewhere between 500,000 and 600,000 Reddit members of our sub-forum. Essentially they are regularly posting, whether it be in races, in between races, pre-season, post-season. They’re posting lots of stuff, whether it be complaints, happy things, good things, bad things – they don’t just complain the whole time; they do give some quite insightful analyses.”
Here Roberts and his team apply linguistic analysis, which is merged with minute-by-minute time codes linked to video replays, to understand how fans react to a given situation, such as a penalty decision, an overtaking move or whatever. This information is in turn fed to Symonds and his team “who take apart the data to understand what drives race engagement.”
Using charts, Roberts explains some of the findings: “The races that are the most incidents-packed get bigger spikes. So obviously Azerbaijan 2017 [which featured a series of dramatic incidents including Sebastian Vettel’s deliberate swerve into Lewis Hamilton] was huge in terms of the amount of traffic and people talking about those races.
“The spikes usually are around controversial results or in-race decisions. We’ve looked at this data compared to some of the other tools that we have. We do other research projects, and we did a Galvanic Skin Response [more on that below] and the data really matches up.”
Advert | Become a RaceFans supporter and
The plan is to roll out these analyses to further social platforms and venture beyond English language platforms, and eventually extending the analyses to every race.
“We appreciate it’s early days yet, but as seasons go by we’ll build up that robust sample of what makes a great race, what makes a bad race, and that’s essentially where we want to get to. And then help inform decisions on the back of that. It’s being used quite heavily by the broadcasting team as well, because obviously they can see what they’re doing right and what they’re doing wrong.”
Do they guide conversations by ‘planting’ statement or questions? The answer is an emphatic “No”. However Symonds points out it is “difficult” to frame questions correctly without “leading the witness”, as he calls it.
“Occasionally we have to add some explanation [in polls]. We were talking about alternatives to penalties earlier. If you just put a question out there saying, ‘Do you think a ballast penalty is better than a grid penalty?’ it’s really difficult for them to understand all the implications of it.
“So we do put a little bit of a narrative around the arguments. We try not to lead the witness. The unintended consequences of any regulation change are actually quite difficult even for us to determine, and for even the avid fan. In fact, avid fans are sometimes actually worse than casual fans, because they’re so pre-programmed.
The correct framing of questions is crucial, says Roberts, using research into possibly sprint races as example: “About a year-and-a-half ago we just tested the waters around sprint races, and the way we asked questions really, really annoyed the respondents.
“They felt we were leading them to say that we were definitely going to have a sprint race next season, and that we weren’t asking their opinion properly.”
To supplement those conversations, Roberts and his team have begun using Galvanic Skin Response studies of small numbers of fans. This involves physically monitoring their skin using sensors which track activity in the sweat glands to assess their arousal to a particular stimulus – in this case, the events of a live race.
Go ad-free for just £1 per month
“We did a pilot for this back in Silverstone in 2018, and then this season we’ve done it at every race and next season we’ll be doing it at every race,” says Roberts.
“We wanted to understand the race experience. We wanted to understand [fans’] on-track highs and lows, and then what content essentially makes a good race, what makes a bad race, and which races have been better than others.
“The reason why we did this particular methodology is because if you ask people questions, you tend to get sort of rational responses, whereas 90% of human behaviour is driven by emotion.
“One example is, this essentially tells us when people are engaged, because we know by their skin response to various incidents within the race. It was in the  British Grand Prix, on this people were not that engaged when Hamilton went from sixth to third because two cars went in the pits.
“But on the survey we did alongside, where we asked people, ‘Are you happy or are you not happy?’ they all said they were really happy when he went from sixth to third.
“In reality it wasn’t that exciting because nothing happened. He just drove out on track and he just went up a few positions. So that’s just an example of how people will answer a survey differently from when you’re actually trying to measure their response to something.”
Symonds points out F1 is the first sporting activity to use this technique in fans’ homes – in other words, in a television viewer’s natural viewing environment. This means respondents are not required to head for a laboratory or don headsets or tracking glasses.
F1 recruits 100 representative people per race for biometric analysis paying them £50 per race in return for strapping a sensor to the palm of a hand. Currently all respondents are UK-based and recruited using selection panels, but the plan is to roll it out to other territories.
“They just stick it onto their hand. They can do it all in-home, they don’t need to be plugged in or anything, and they just literally sit and watch the race.
“We’ve signed this [technique] up again for next year, and hopefully for a number of years after that, that numbers will go up much bigger, and we can then start looking at sub-groups. But it’s just more a function of costs, and it takes quite a lot to get this project set up. So we reckon 100 is a manageable number at the moment,” says Roberts.
Using a time-coded chart he illustrates last year’s British round: “The race start there, a safety car tends to dip down, people get a bit less engaged. That’s the high towards the end of the race where people come back in again and then you’re starting to see that drop-off.
“We’ve done a lot of work around this and with TV ratings, and we have about an 11-minute window to keep fans engaged: 11 minutes is the most likely time that casual fans will switch over. So they watch the start and then it’s clear who’s in the lead and they will switch over. Then they’ll come back at the end.”
Every sport is different, says Roberts, using football as example: “If the team is four-nil up at half time, suddenly it just drops off. So if we can make the racing more competitive and more exciting, then we’ll see that line go up rather than see that dip before they come back again.
“That’s our goal; it’s to have that line go up rather than see the dip that we do.”
Is the trick not, though, to achieve an upward trajectory organically rather than artificially through devices such as DRS, or worse, gimmicks?
“[That’s] right,” says Symonds. “If there are five groups of two cars battling, it doesn’t actually make any difference, it doesn’t get any better. All we need is to get two cars battling.
“So you’ll say, ‘Well, what can we do about that?’ And you can’t artificially make two cars battle. What you can do is try and get the performance of all the cars more equal, and then you’ve got a much higher chance of a battle happening. So that’s one thing I think that we can learn from it.
“We don’t want to make it artificial. What we want to increase the jeopardy. At the moment everything is so damn processional in Formula One; if we get a little bit more jeopardy into it, a little bit more of these sort of things happening that the fans seem to enjoy so much.”
Can you have too much of a good thing? July’s chaotic German Grand Prix may have been just that, says Roberts. “The casual fans were actually more engaged than the avid fans, because they loved the jeopardy of it, whereas the avid fans probably found that it was just a bit too much for them. We’re finding some quite interesting nuances.”
Equally, from viewer reactions F1 has learned that onboard cameras don’t maintain viewer attention unless there’s another car ahead, while drops in attention can be countered through replays during lulls in action. Biometric data can also be used to track the individual impact of commentators, but Roberts refuses to be drawn on this point…
One of the challenges, Symonds emphasises, is that tracking is only relevant if a particular incident is broadcast: “If the director misses something – we may see [something] in our data there’s something really interesting going on and we’d expect to see a high response, but if it’s not on TV, we won’t see that response. So we have to apply that filter when we do our analysis as well.”
According to Broadcasters Audience Research Board data, avid F1 fans watch 16 or so races per year, while casual fans consume a quarter that. Hence that is F1’s target market for conversion to ‘avid’ could quadruple audiences, with such data being crucial F1’s future direction.
Clearly, after just three years under Liberty, it is still early days for F1’s market research to play out fully, but the point is that science is at last being applied to this most scientific of sports, and that can only stand F1 in good stead as it faces arguably the biggest challenges in its 70-year history.
Crucially, though, the data should be used to inform decisions, not to determine them. As Pat Symonds, whose task it is implement change, “That is important.”
The mere fact that F1 recognises this provides hope for its future.
- The year of sprints, ‘the show’ – and rising stock: A political review of the 2021 F1 season
- The problems of perception the FIA must address after the Abu Dhabi row
- Why the budget cap could be F1’s next battleground between Mercedes and Red Bull
- Todt defied expectations as president – now he plans to “disappear” from FIA
- Sir Frank Williams: A personal appreciation of a true racer