TL;DR
- Refreshing an ad placement 1 time reduces CPM by almost 10% (CPM = cost per thousand impressions)
- Refreshing an ad placement 5 times reduces CPM by 25%
- Using refresh increases revenue-per-page for long page view times
- Using refresh could decrease site performance
- Using refresh could increase data usage
Many techniques are used to draw more value out of an impression on a web page. These can vary from placing more ads to the page, to adding higher paying units (like video), making ads larger, or placing them in front of the page’s contents and forcing interaction. Another technique that can be used is the practice of refreshing, or reloading, the ads that are already displayed on the page.
The reasoning behind implementing ad refresh is fairly simple: if the first page load resulted in 3 ad impressions, and the user remains on the page for a long enough period of time, refreshing the ads can increase the number of ads shown to the user from 3 to 6, and should produce higher revenue-per-session for the publisher.
This sounds like a pretty straight-forward process, however, are the impressions on the refresh worth as much as the initial load? Is there a decreasing return for the cost of serving new ads to the same impression? Is implementing ad refresh going to have a meaningful impact on revenue or user experience?
Luckily we have some data on how refresh can affect CPM for ads.
Using ad refresh can be an effective way to increase page revenue, but only for specific use cases. For example, if a page operates by changing the content in response to user actions, the effective time on page will be high. These interactions can also be used to ensure that the refreshed ads remain in view ” since the publisher knows where the user action occurred, they also know where the ads will be in relation to the viewport when the ads are refreshed. In this case, using refresh could result in greater revenue per page.
For the vast majority of web sites, the time on page is short enough that refresh would never be triggered. The cost of implementing and managing ad refresh would outweigh the benefits of using it. Advertisers also generally dislike refresh: the CPM they agree to pay is based on the assumption that the ad exists on the page for the duration of the page view ” regardless of how long that page view is. If the ad slot for that page view is now limited to 30 or 60 seconds, the price of the ad slot should probably change accordingly. For a publisher to keep the best relationships with its advertisers, avoiding the use of ad refresh would be the safest long-term approach to creating ad revenue.
In addition to these points, if ” along with the other challenges of implementing ad refresh ” users are annoyed by the constant refreshing and therefore do not return to the site, this will have a much larger and lasting impact on revenue than any of the technical details already mentioned.
In general, for the majority of publishers, the downsides of using ad refresh outweigh the upsides.