As long as banner ads have been on websites, the number of clicks they garner has been the most important performance metric for an ad. Ads that get fewer clicks are canned in favor of those that get more.
I questioned this (with Seth Godin’s unwitting help) a few years ago in College Branding and Banner Ads, and described both public and private research on brand perception changes from ads that people didn’t interact with. Now, there’s new research that underscores the weakness of the link between ad clicks and conversions (orders or requests for information).
Being Seen Beats Being Clicked
Comscore, the web stats aggregator, and Pretarget, an “intent targeting” company, ran a large-scale study involving 263 million ad impressions. They looked at conversion events (such as filling out a form or downloading software) vs. how users experienced the display ad. According to Pretarget Founder Keith Pieper, “Your ad being seen matters more than your ad being clicked – if you have a back-end conversion metric.” A release from the two firms says,
The results showed that ad hover/interaction (correlation = 0.49) and viewable impressions (correlation = 0.35) had highest correlation with conversion, while gross impressions (correlation = 0.17) was significantly lower. Perhaps most interestingly, clicks (correlation = 0.01) had the lowest correlation with conversion, far under-performing all other metrics analyzed in the study. These findings suggest that advertisers and media planners ought to break their addiction to clicks and instead look to more meaningful metrics for evaluating campaign performance.
It may seem too obvious to mention, but ad visibility is a huge factor in conversion (and, presumably, brand lift as well). The release cites a Casale study that found conversion for “above the fold” ads (i.e., ads viewable without scrolling) was 6.7x higher than for below the fold ads.
The viewability metric is important because traditional ad impression counting includes ads that are included when a page loads in a browser but which the user doesn’t actually see. Conversion had twice the correlation with the number of viewable ad impressions in the study vs. total ad impressions.
The most telling metrics in the study were interaction and hover – they had a 0.45 correlation with conversion (compared to a minuscule 0.01 for clicks). This seems counterintuitive, but makes sense if you assume that the low correlation for clicks isn’t due to some negative aspect of the post-click experience but rather that the number of clicks is so low that it explains very few of the conversion events.
Avoid Bad Impressions
An area the the study didn’t look at was overall user experience, but we all know that this varies greatly between sites. Avinash Kaushik highlighted this issue the other day in his post, You Are What You Measure, So Choose Your KPIs (Incentives) Wisely!. He compared the user experience of sites that seek to maximize ad impressions with content that requires reloading a page every few seconds (e.g., Yahoo News photo slideshows) vs. sites that create a far better user experience by avoiding needless refreshes and preloading images for blazing-fast viewing. Of course, as Kaushik notes, pageviews are often easier to measure than more meaningful indicators.
Kaushik was focused on the misuse of pageviews as a metric for user engagement, but there’s a lesson for advertisers, too. Clearly, an ad impression on a slide in the middle of a slideshow that annoys users as they try to click through as quickly as possible isn’t as good as an impression on a page which encourages viewers to linger and enjoy the content.
Takeaways for Advertisers
Some ad media, like pay-per-click search ads, require a focus on clicks and conversion so that ad campaigns can be optimized and elements with negative ROI can be eliminated. Display ads, though, are a somewhat different situation. There may be multiple exposures and a variety of steps that lead up to a conversion event, and evaluating ad performance purely in terms of clicks (and related metrics like click through rate, cost per click, etc.) may produce misleading guidance.
Choose Your Metrics. It’s clear that getting above the fold is truly critical. If not above the fold, then the ad should be in an area where most users will scroll and be exposed to it. Attention-getting ads that cause the user to interact or hover without clicking will out-perform those that don’t; these metrics are measured less frequently because doing so requires non-standard technology. Continuous measurement may not be practical in every case, but periodic testing of different ads would at least point the advertiser in the right direction.
Brand Lift. This study didn’t look at the branding aspects of non-clicked ads, but we know from the work cited in my earlier post that even if no click occurs and no conversion event is recorded, brand familiarity and perception can be improved by exposure to banner ads. The message remains the same: think beyond the click!