The Data Geek’s MMQB
Some of you may recall our analysis of consumer attention during last year’s Super Bowl. There have been a great number of follow-up studies over recent days, including a Business Insider post on Twitter buzz during and after the game, according to which Netflix’s “Stranger Things” was the game’s most popular ad.
At Dstillery, our data analysis takes a different tack. We operate an intelligence platform that observes 100 billion U.S. consumer events per day, across the full array of digital devices. So how do we apply that intelligence to an analysis of what commanded viewers’ attention during the Super Bowl broadcast? To quote Christopher Zara in his Fast Company coverage of our analysis from last year:
“It’s all thanks to the wonderful world of programmatic advertising. When 111 million people are watching a single TV broadcast at the same time, a sizable chunk of them are on their phones or computers, too. As they watch the big game, their attention is divided between their TV screens and whatever distraction is beckoning online, whether it’s Twitter, Candy Crush level 15, or what have you.”
So, yes – we watched closely again this year, as consumers were engrossed in the annual spectacle! Let’s take a look at the game from a mobile device user perspective. The graph below shows the relative reduction in bids from mobile devices compared to an average Sunday evening.
At 6 pm ET, it is all still business as usual with the bids arriving at par with an average Sunday. But the excitement is slowly picking up. By the National Anthem, we already observe a 15% reduction in bids. The first clear highlight is the coin toss at 6:33 pm with a reduction of 20% in mobile advertising auctions compared to the same minute on an average Sunday.
From there, the attention wanes a bit, with the occasional blip at the Blount fumble and for each Atlanta touchdown, particularly the pick six at 7:43 pm that made it 21 to 0. By 8:05 pm, after the Patriots fail to get in the end zone, people’s attention span for either the game or ads is clearly exhausted – in fact, attention drops below normal Sunday levels.
And then the halftime show starts at 8:13 pm – and we see a good 13 minutes of eyeballs glued to the screen at a record high of 25% above typical levels. While Beyoncé’s performance from 2016 (including tripping on stage) was the clear popular favorite during the broadcast, it did not match the dramatic command of viewers’ attention that Lady Gaga’s acrobatic performance exhibited!
The remainder of the game proceeds with the expected pattern of attention spikes at touchdowns and relative tune-outs during the ad breaks. The overtime touchdown is only the third most attention-commanding event, below the halftime show and the first coin toss.
From a marketer’s perspective, we see a notable pattern repeating from last year. Despite the huge rally by the Patriots and commensurate spikes in attention at critical second half game events, the average attention to ads dropped significantly in the second half. In general, people are watching through the ad breaks during the first half, with only modest interruptions in attention. But they abandon their viewing and return to their phones during the second half commercial breaks.
A slightly different picture emerges for different device types. The plot below shows the same analysis, but this time adding tablets and desktops.
All device types show a consistent reduction in ad-monetized activities during the game broadcast. The impact of the broadcast is most pronounced for desktop, where the reduction in attention to online ads is consistently 20%. Halftime still stands out, but the final touchdown reaches an equivalent peak of about 27%.
In addition, we uncovered a small but revealing detail: desktop users cared relatively more about the anthem than the coin toss and even more so about the early events (the first two touchdowns). Only after Atlanta built their big lead did the computers start to come back online, receiving ads. For the entire game, the relative attention remains above 15% until people finally go to bed.
If we had to make a conjecture, marketers probably had a higher impact on desktop owners than on the ‘generation mobile’ with their fickle attention and practice in multi-tasking. Given how much brands spend on ads during the Super Bowl, the question of who is paying attention becomes quite relevant. Good creative design is difficult in the best of cases, but creating the right spot for the Super Bowl is fraught with challenges.
A recent post-Super Bowl article from Cards Against Humanity shares some brutally honest reflections on lessons learned trying – and failing – to create the right spot for the pre-game coverage:
“ While we succeeded creatively, the advertisement showed a disappointing return on investment ($0), and we are now going out of business.”
Having come to ad tech relatively recently, I arrived in this world of “Mad Men” with a strong attachment to the scientific method, experimentation and data. The conventional practice of creative design and random selection often leaves me with a mixture of bafflement and dismay. This is vividly echoed in the lessons discovered by the Cards Against Humanity team:
“Instead of asking one person if the ad was effective 70 times, we should have asked 70 people one time [each].”
One of the upsides of a programmatic environment is the ability to not only see your customer’s behaviors as they go about their everyday lives, but to run actual tests and observe the effectiveness of your ad in near real time at unprecedented scale. While the Super Bowl will continue to be a mass spectacle attracting huge attention from both marketers and consumers, the future of impactful marketing is rooted in a meaningful collaboration between creativity and data science.