Why Clickthrough is a Terrible Engagement KPI

Dwell on Branding Research Report, April 2010 (Microsoft Advertising)

Dwell on Branding Research Report, April 2010 (Microsoft Advertising)


Clickthrough used to be king. Now it's barely a squire.

Display advertising is about delivering value to a brand, whether it means increasing awareness or perception, or nudging consumers further along the purchase funnel. However, the predominant metric in digital advertising continues to be clickthrough rate (CTR), even though it's an unimportant measure of a display ad's effectiveness.

Come again?

I'll repeat: CTR is an unimportant measure of a display ad's effectiveness.

This isn't a matter of opinion.

  1. Only a small segment of the overall online audience actually clicks on display ads. According to a comScore study, 8% of Internet users account for 85% of all display ad clicks, and 72% are uncomfortable clicking on ads altogether. By measuring clicks, you ignore the impact an ad has on almost three-quarters of your audience.
  2. Those that do click a lot are an unattractive audience. A study conducted on behalf of Starcom and Tacoda showed that "heavy clickers," those 6% of Internet users that constitute 50% of all ad clicks, skew younger (25-44) and less affluent (household incomes under $40,000).
  3. Clicks don't correlate to conversion. Whether a conversion means a purchase or a request for more information, ad clicks correlate remarkably poorly (r = 0.01), according to a study commissioned by comScore and Pretarget.

If clicks are a lousy measure of a display ad's ability to deliver value, then what's better? Research points in the right direction.

The comScore/Pretarget study found the factor that correlated with conversion most strongly was ad interaction or hover (r = 0.49). Research conducted by MediaMind bolsters this conclusion: increasing hover from 5% to 15% led to a 45% increase in conversion.

Similarly, research by Microsoft Advertising found that longer "dwell times" (hover/interaction) correlated with likelihood of branded searches. The study looked at 800 campaigns, and segregated results based on how long on average users spent engaging with the ad. Those campaigns in the top 10% with respect to dwell time spurred more than three times (39%) the branded searches of those in the bottom decile (12%).

A case study prepared by Microsoft Advertising looked at a technology advertiser with high-engagement ad creatives (homepage takeover, expansion, wallpaper) and high dwell time (50% higher than the tech vertical average). The results were impressive:

  • Branded search: 14% more likely
  • Site visits: 3.75x more likely
  • Ad recall: 61%
  • Brand awareness: 7 basis-point increase (from 40% to 47%)
  • Brand favorability: increased from 15% to 34%
  • Purchase intent: 3 basis-point increase (from 67% to 70%)

So what does matter? Metrics like hover time, interaction rate, and engagement rate are far better indicators of how useful or entertaining an ad is, and the more a user interacts with an ad, the more likely it will leave a lasting—and positive—impression.