Simultaneously served ad impressions is a side effect of RTB auctions which raises a lot of eyebrows with advertisers. Can there be any value in showing a consumer two or three of your ads all at once or is it a waste of budget? Simultaneously served impressions are an uncontrollable symptom of RTB auctions, where a separate auction is held for every as-yet-unsold placement on a webpage. When an advertiser is targeting a specific high-value user with a high bid, it’s inevitable that sometimes they will win more than one of those auctions, resulting in several of the same ads appearing on the page. However, is it just a wasted opportunity for the ad to be shown to the person at another time, or to someone else, or could there actually be value in these impressions? We’ve put our data science team on the case!
How do we measure simultaneous impressions? More detail on zero delivery time
Multiple ad impressions on the same page are measured via delivery time. We define delivery time as the difference in time between ads in a campaign being shown to a unique user. Therefore, we determine zero delivery time as when there is no difference between when one or more ads are served, meaning one or more ads from the same campaign are shown simultaneously, rather than displayed sequentially over a longer period. For example, as a consumer visiting a webpage, you might see two or three ads from an advertisers’ campaign – often the same creative – on the page at the same time.
How do we know that zero delivery time impressions occur and how do we calculate it? Click here to know more
What is the effect of simultaneous impressions on CTR?
We chose to look at click-through rate (CTR) rather than conversions so that we have enough data to work with in our sample datasets. We wanted to see how CTR changes with the number of impressions served simultaneously, looking at two impressions served simultaneously (double contact) and three impressions served simultaneously (triple contact) The results of our analysis show that multiple impressions served on the same page can have a positive effect on CTR. We decided to compare this to the impact of a billboard.
Comparing simultaneous impressions to billboards
Impressions with zero delivery time (the most we’ve seen is five at time) can take up a lot of space on a webpage, which led us to draw a link to billboards. The most common banner sizes to appear as simultaneously served impressions, for both double and triple contacts, are 300×250 and 728×90 (IAB standard sizes). Billboard size is 970×250 and are generally more expensive. For sequential impressions (non-simultaneous), we see a CTR after three contacts that is still under a third of that seen for billboards – showing the much greater power of a billboard.
From these results, we go on to compare our zero delivery time impressions to billboards. Two simultaneous impressions don’t show a dramatic uplift compared to sequential impressions, resulting in just over one third of the CTR of the impressive billboard. However, for three simultaneous impressions, we see a big improvement in CTR, jumping to around three times the CTR of sequential impressions, nearly matching that of the billboard.
Finally, we wanted to understand how frequently zero delivery time impressions occur. We calculated this by looking at the unique user and timestamp combinations divided by all user and timestamp combinations. With this calculation, we found that it happens about 5% of the time. Further analysis showed that often poorer quality websites tend to have a high number of simultaneous impressions. These websites can simply be excluded from targeting by the campaign manager.
In summary, we have myth-busted the commonly held belief that zero delivery time impressions are a waste of advertisers’ budgets. In specific sizes, three small simultaneously displayed banners can have the same effect as one billboard, giving CTR a huge uplift. There is currently no way to act on simultaneous impressions, and luckily our findings show that zero delivery time impressions are nothing to worry about. Therefore, it’s best to take peace of mind from this analysis and enjoy the happy side-effects as they happen!
https://www.bannerconnect.net/wp-content/uploads/2016/09/Freq-cap-visual-bw.jpg10801080Tim van Rijthttps://bannerconnect.net/wp-content/uploads/2015/09/logo_bannerconnect_72.pngTim van Rijt2016-09-07 15:24:472018-05-17 11:07:13Here's What Frequency Caps Get Wrong - And How to Improve Them