Third down, fourth quarter, Minnesota Vikings quarterback Kirk Cousins drops back against the blitz and locks onto a receiver while all eyes in the stands at U.S. Bank Stadium watch the tight spiral land in the hands of . . .
Hold up. Not so fast.
Everything you think you know about crowd behavior at U.S. sporting events probably needs a solid reality adjustment.
Yeah, there’s data for that.
The Vikings are among a growing number of teams across a handful of sports that utilize CrowdIQ, an analytics company that uses artificial intelligence (but not facial recognition) to interpret aggregated fan demographics and behavior down to the second.
With offices in Cape Town, South Africa, and in Minneapolis, CrowdIQ formally launched in March as a sister company of FanCam, which began in 2010 with the idea of providing fans with a photographic keepsake and brands with personalized fan engagement. Attendees at a game could review the photos, tag themselves and share it, usually with an associated marketing tag.
After years of reviewing such images, FanCam founder and CEO Tinus le Roux noticed broad trends in crowd compositions: this crowd was younger, that one was more female, another was more attentive to the game.
But attentiveness has its limits.
As far back as 2019, the Vikings discovered that “there was no time when more than 40% of people were looking at the field at any given time when they got in the stadium,” says Vikings director of business analytics Kendall Peters.
Those pivotal plays when the game seems to hang in the balance? That’s when the action attracts only 40% of eyeballs.
“We were kind of floored by that,” Peters says.
Anecdotally, it makes sense. There was a time in sports before smartphones, when all eyes seemed to be glued to every move that the likes of Michael Jordan made, and after smartphones, when fans don’t even see balls coming at them in the stands at dangerous levels of speed (see the embedded tweet below). Even modern stadiums are built with the idea of creating an immersive experience that gives fans several options to explore beyond the game that’s actually being played by world-class athletes.
Have you ever been THIS into your phone?pic.twitter.com/11gkpNjvNu
— Danny Zederman (@DZederman) June 21, 2021
Beyond the broad strokes, there’s a sea of granular data that can influence team and league operations and game-day experiences. Says Peters: “There’s so much behavioral information that we’re just scratching the surface of having a chance to test further.”
The Vikings have found that the average amount of eyeballs actually watching the action at any given time is closer to 30%. There are a multitude of factors that influence that number—just as there are other metrics about fan attention that range from sponsorship activations to videoboard entertainment to concessions management.
Using the same 360-degree gigapixel photography as FanCam, CrowdIQ helps venue operators understand who is attending an event and what forms of entertainment are drawing their attention. The company now has 13 clients, primarily focused in the U.S., including the NFL’s Vikings and Jaguars. (To date, CrowdIQ has only piloted attention tracking with the Vikings.)
“We have per-second analysis of where everyone is looking. If you tie that up with what’s happening on the big screen, you certainly know which of the sponsored assets are performing better or worse,” le Roux says. “Now we have the data to actually put science against them saying, ‘This is what’s really happening, and this is how you structure your in-game experience so that it’s perfect.’ ”
What CrowdIQ does not do—a point emphasized by both le Roux and Minnesota Vikings executives—is facial recognition. There is no CrowdIQ database of people for the AI system to compare crowd images to; thus there is no identifying one-to-one correspondence. CrowdIQ instead compares images with anonymous male and female faces of various ages. “I want to be clear that we don’t do [facial recognition]. In fact, we stand quite against it,” le Roux says, adding: “It’s as dumb as the system is. That’s a benefit of AI, is you can make it a dumb system.”
Although mobile ticketing has helped a number of teams learn more about who is actually entering stadiums for games, that information is never complete. The Vikings use CrowdIQ to fill in the gaps or validate the data they have. “For some of those unidentified people sitting in the seats, at an aggregated level, it’s really good for us to also know the age, the demographics, where they’re sitting, where they’re looking and all that,” Peters says.
That has led to a better understanding of crowd dynamics. There’s a clear delineation between someone who attends a game with a noon Central kickoff as opposed to a 7:30 p.m. game on national TV—and when they actually get to their seats. Peters notes that CrowdIQ photos provide data points on the typically late-arriving primetime crowds.
“In some of those night time games, we saw a gender shift,” says Genette Sekse Amar, the Vikings’ director of business strategy. “For the next few years now, we’re able to plan out: what does the presentation look like? ‘OK, it’s going to be a night game, we see higher skewed female attendees and a little bit younger.’ It gave us a bit of a better insight on those because, from a ticketing standpoint, we didn’t see a lot of difference in the data.
“We used fan season-ticket data and were like, ‘OK, we should play old rock. That’s what we played in the past,’ ” she adds. “But then we’re like, ‘Oh, for these primetime games, we see the demographic is [younger] so we should probably shift our music or shift our entertainment value.’ ”
The analytics can help inform decisions not just on the macro level—such as music—but even about specific moments. The Vikings mused about using that second-by-second analysis that le Roux mentioned to improve sponsorship implementations. “We can layer that data with our partners that are spending money with us in-game to enhance and make these promotions better,” Sekse Amar says.
CrowdIQ wants to take it a step further, with le Roux citing the “big screen fatigue” that comes with oversaturated video programming. He says sports teams could explore dynamic pricing for time slots on a stadium Jumbotron.
“Our job has been to get the data. We’re at a point now where we want the conversation on the industry side to be, what do we do with the data?” le Roux says, adding: “It’s never been possible to quantify crowds this way. The technology didn’t exist. Now that it exists, you can do it at scale.
“We catch it multiple times during a game, so we can tell you how fans follow your seats, how they leave, and all that data. The interesting opportunity now is for the sports industry to use this data to actually go and get more sponsors.”
Question? Comment? Story idea? Let us know at [email protected]
mindtalks.ai ™ – mindtalks is a patented non-intrusive survey methodology that delivers immediate insights through non-intrusively posted questions on content websites (web publishers), mobile applications, and advertisements (ads). The conversation is just beginning !, click here to sign-up and connect with other mindtalkers who contribute unique insights and quality answers on this ai-picked talk.