• by Tobi Elkin , Staff Writer @tobielkin, February 7, 2017

Just in case you didn’t process enough information from Super Bowl LI, ad tech firm SourceKnowledge analyzed real-time data on the game, comparing this year’s numbers to data from 2016 and 2015. The analysis made for some interesting comparisons about the real-time behaviors of consumers across devices and interest categories.

The SourceKnowledge dataset consisted of 7,357,628,837 rows of data from Super Bowl XLIX, Super Bowl 50 and Super Bowl LI. Each row of data represented an ad opportunity that the firm’s bidder analyzed. 

First off, SourceKnowledge analyzed traffic patterns to find that overall traffic volumes across all devices, interest categories, and inventory sources decreased by 4% from 2015 to 2016, but increased 12% from 2016 to 2017.

While overall traffic decreased, mobile traffic actually increased by 18% from 2015 to 2016, a sizeable jump. Mobile traffic increased just 4% from 2016 to 2017,

SourceKnowledge attributed the increases and decreases to two main trends: first, the growth in mobile content over the last several years. Second, the nature of mobile—it's considered a complementary channel that can be used in a social or shared context while consuming content on a principal screen (TV).

SourceKnowledge found that Web traffic decreased compared to the prior  two weeksleading up to the Super Bowl games between 6 p.m. to 10 p.m.  by 28% in 2015, 20% in 2016, and 24% in 2017. The idea there was to get a baseline for normal traffic patterns that could be used to compare against the data from the Super Bowl  game itself.

However, in 2015 mobile traffic only decreased by 3% during these hours, while desktop traffic dipped by slightly more than 36%. That trend was similar in 2016—mobile traffic actually increased by 1% during these hours, while desktop traffic dipped 42%. This trend continues in 2017, as desktop traffic decreased 29%, while mobile increased 4%.

In the period between 10 p.m. and 12 a.m. on Sunday, mobile traffic increased compared to previous weeks by 9% in 2015, 15% in 2016, and 12% in 2017. Similar patterns held true with the bulk of the increase within mobile for all years. What does SourceKnowledge attribute these patterns to?

“Mobile is a complementary channel to traditional television. Unlike desktop devices, mobile can be used within a social context and outside of the user's home without much friction,” Patrick Hopf, SourceKnowledge president and co-founder, told RTBlog via email. “This data demonstrates that while desktops compete against big television events, mobile does not. In fact, large television events like the Super Bowl can cause users to increase their mobile usage.”

 In terms of cost and competition, SourceKnowledge found that competition increased from 2015 to 2016 as the effective CPM across all devices, interest categories, and inventory sources increased by 23% from 2015 to 2016, and jumped up 44% from 2016 to 2017.

“Competition for consumers’ attention during this time is increasing year-over-year. Advertisers are shifting their ad spend to digital, and it shows,”  Hopf said. “They don't need to commit to a single multi-million dollar ad spot. Digital also yields significantly better addressability and attribution possibilities.”

When it comes to consumer behavior, viewers were almost equally likely to click on ads: a 1% increase in click-through rate from 2015 to 2016, and 0.4% from 2016 to 2017.

While ad opportunities were concentrated in the news and sports verticals for most of game day in 2015, 2016, and 2017, shopping-specific domains saw an increase in ad opportunities between 10 p.m. and 12 a.m. on game night, a 6% increase compared to the rest of the day in 2015 and an 11% increase in 2016. And in 2017, shopping-specific domains saw a 2% increase from 2016 to 2017, while auto represented the breakout category in 2017, increasing by 8% from 2016 to 2017.

In terms of consumer behavior, Hopf said: “While consumers' likelihood of clicking on an ad doesn't change dramatically, their browsing habits do. The majority of consumers generally react to advertising not by clicking and purchasing, but by researching the products they’ve seen, comparing prices or alternatives, or reading reviews. It's important for advertisers to track every touchpoint within the conversion funnel and focus on measuring impact instead of optimizing for clicks.”