“Ads don’t work if no one pays attention to them. Duh.” Ok yeah, this is obvious. But do ads work when people actually do pay attention?

It turns out this is a surprisingly difficult question to answer. Also, the answer is yes — under certain conditions.

Why it’s difficult: over 99 per cent of ads are not paid attention to, so it’s not possible to measure the impact of ads without isolating the ads people actually pay attention to. Otherwise the findings would be at least 99 per cent meaningless.

Wait, what?

The average US internet user is being served well over 1000 digital ads every day, 1147 in programmatic alone. Quick maths: how many ads do you remember paying attention to yesterday? Right — less than one per cent.

So this means we needed to isolate the ads people actually pay attention to if we wanted to draw any conclusions from the data.

How? After all, you can’t get millions of normal people to let you watch them interact with ads. Anyone who would let you observe them while interacting with ads is not normal. These sorts of random volunteers can work when you’re talking about user testing new apps and such — tests that are optimising for ease of use, intuitive interfaces, common confusions, etc. But persuasion is different, more subtle. You need a blind study to measure persuasion.

We knew we could use Moat to measure the amount of time users spent with an ad unit, we also knew users usually have ads on the screen while they’re consuming other content. This meant metrics like Time on Screen (ToS) would be meaningless unless we were taking up the entire screen.

But we didn’t want to use something like an interstitial, the fancy cousin of the pop up. Blocking a user’s journey without their permission meant that any gain to awareness would come at the risk of negative brand perception.

So we used a full screen mobile creative that took up the entire screen on a user’s device. But importantly, we also gave the user choice about whether or not to keep it on the screen. The user could simply swipe the ad away at any time if they weren’t interested. This ensured both attention as well as intention.

We built custom creatives for 65 brand clients ranging from Land Rover and BMW to Microsoft, Wells Fargo, Sephora, Adidas and 59 more. These were all brands the majority of consumers have heard of before, which makes creating lift that much more difficult. Often we used their existing assets but we typically designed the overall experience to drive the greatest impact to the stated goal — awareness, recall, consideration, etc. Moat was embedded in the units to verify humanness, viewability and time spent.

We used third-party vendors, Millward Brown and Lucid, to survey users who had been exposed to these ads, asking them the normal types of questions brands use these surveys for. They separately surveyed a control group of users who had not been exposed.

We then further segmented the exposed group according to the amount of time each user spent with the ad. With enough data we were able to compile an understanding of the impact of attention on brand lift — and the ability of incremental time to generate incremental lift.

Aggregating this much data across 65 studies over two years was a major undertaking, but the significance of this amount of data can’t be ignored.


Brand Lift as a Function of Attention Duration

It works!

Here are some key findings:

  • Each incremental second of attention generated 1.65 per cent relative lift in consideration and 1.81 per cent relative lift in awareness.
  • Gains in awareness were most noticeable in the first five seconds of attention and tapered with additional time.
  • Consideration had a more linear correlation to time spent — almost perfectly linear. This rings true — the more time I spend examining a product in a store, the more likely I am to consider buying it.

But there are some major caveats:

  1. Time spent is meaningless without controlling for attention. Otherwise, the data would be at least 99 per cent meaningless.
  2. Attention is really hard to measure at scale, and it helps to use a creative unit that takes up the whole screen so you know that’s what the user is focusing on.
  3. Don’t take up the whole screen unless you make it really, really easy for a user to scroll the ad away. You might not care, but your prospective customers do.
  4. Creative matters a tonne. Across 30 brand studies, the incremental lift per second of attention was very similar, but overall campaign lift varied based on how well the brand message aligned with the goal.

Dave Lando is the Marketing Director of Parsec Media

Image credit: Photo by Joshua Earle

Previous post

Exposed Facebook Data of 533 Million Users Re-emerges Online

Next post

Minicast: How Ethical Is Your AI?