Attention metrics have mainly been used to measure lift in upper-funnel KPIs, like brand awareness.
But advertisers are starting to gravitate toward attention as a way to measure a campaign’s impact on the lower funnel, including conversion rate.
Luxury car manufacturer Audi recently experimented with using attention metrics to algorithmically adjust programmatic bidding on ad inventory based on the amount of attention those ad placements are likely to draw.
Audi then measured the impact high-attention placements had on post-click conversions.
The idea for the experiment came from a desire to understand how certain ad placements that aren’t highly valued from a viewability standpoint can nonetheless impact lower-funnel campaign performance, said Filip Pujic, who joined Audi as a digital marketing team leader in February after a stint at MediaCom Switzerland, Audi’s agency.
The experiment involved an online display campaign that promoted Audi’s electric vehicle offerings. The campaign, which ran from June to July, focused on impressions served to users in Switzerland.
Audi worked with MediaCom Switzerland, attention startup Adelaide and GroupM’s Xaxis to design a programmatic bidding algorithm that would place higher bids for inventory whose attention score reached above a certain threshold.
Adelaide has a proprietary attention metric called AU (which stands for “attention unit”) to determine which ad placements are likely to attract the most attention from potential customers.
Audi activated Adelaide’s tag-based analytics in Xandr’s DSP to measure the AU score for available ad placements, which ran across open web programmatic and PMP-based inventory. The PMP inventory included existing PMPs set up by Xaxis for past Audi campaigns. (No AU-focused PMPs were created specifically for this campaign.)
The custom bidding algorithm would increase bids for media with higher attention scores and reduce bids for media with lower attention scores in real time.
The campaign didn’t use any targeting parameters other than the ad inventory’s AU score.
When the campaign closed, Audi measured the post-click conversion rate for the ad placements it won, but Pujic declined to specify which types of post-click conversions the brand measured.
Audi then A/B tested the results of its custom bidding algorithm campaign against Xaxis’s default bidding algorithm.
The custom algorithm generated a 69% higher conversion rate on open exchange inventory than the default bidding algorithm. Overall programmatic conversions increased by 60%.
The results were a lot higher than the typical conversion rate lift Adelaide sees when optimizing toward high-AU placements, said Marc Guldimann, Adelaide’s founder and CEO. “Typically, we see about 20% to 40% improvement,” he said, “so 69% is exceptional.”
While the results were encouraging from Audi’s perspective, the brand isn’t 100% sold on attention. Audi needs more data to determine just how far attention metrics can drive incremental growth in its lower-funnel KPIs, Pujic said.
“The next steps are to integrate AU measurement in all of our programmatic campaigns,” he said. “It’s necessary for us to have a broader set of data to analyze and to see if the results can be [reproduced] on a consistent basis.”
But given these campaign results, Audi is already thinking that its post-third-party-cookie plans won’t have to rely exclusively on first-party data for targeting lower down in the funnel, which is how Audi has done things historically, Pujic said.
“These [early] indicators tell us that optimization toward attention is a good substitute for the kinds of data we used before,” he said. “I’m not comfortable saying it’s a paradigm shift just yet, but it feels like a change in perspective on how we buy digital media.”
