The Impact of Viewability on Ad Revenue

Posted by Asmaa Bentahar on

Over the last few years, Viewability has become a key performance indicator in the ad tech industry. For a long time, it was considered as a buy-side metric: Advertisers want to make sure that their ads are seen and that their campaigns are performing as expected. As Viewability has an impact on ad performance, it is a metric that should be closely monitored.

To quantify this impact, our experts at Pubstack ran a study to specify how Viewability affects Publisher’s Ad Revenue.


But, what exactly is Viewability?

Viewability is a metric measuring the share of impressions that were actually displayed on the screen of the user.  The Media Rating Council (MRC) introduced the notion of viewability and defined guidelines for the Ad Tech industry. According to them, for an ad to be considered as viewable, 50% of the said ad creative should be visible for at least one second in the viewable space of the browser.

However, other actors consider that these guidelines are not enough and prefer to provide their own viewability definition. For example stating 100% of the pixels must be in view for an impression to be viewable. As a consequence, all publishers (and the whole market in general) are suffering from discrepancies when measuring viewability.


Why should you care about Viewability as a Publisher?

Viewability is an inventory quality metric, therefore, it helps advertisers assess the quality of the impressions they buy from publishers and understand what share of the impressions they bought was actually visible. 

This means that SSPs and Advertisers are using the viewability metric to better assess the value of an inventory or an impression, and therefore adjusting their bid values. Some advertisers can even decide not to bid for a display opportunity if they estimate the expected viewability is too low or even to manually blacklist a poorly performing inventory. For instance, Google AdX or Criteo include a viewability prediction in their bidding algorithms. 

Then, for a poorly viewable inventory, a Publisher will mainly see two impacts:

  • Fewer eligible campaigns for less visible inventory, therefore less competition, impacting  the fill rate and eCPM,
  • Lower bids on less visible inventory, thus impacting fill rate (ability to pass floors) and eCPM.


 So, to what extent does Viewability impact the Ad Revenue?

In Pubstack, we decided to run a two-part analysis to understand this.

  1. Impact of Predicted Viewability on Ad Revenue
  2. Impact of Actual Viewability on Ad Revenue

First, we decide to measure the Impact of Predicted Viewability on Ad Revenue. The idea behind this is that SSPs/DSPs only get predicted viewability info at the bidding time (either at auction level or based on historical data). This means that they don’t really know whether the opportunity they bid for will be viewable or not. Also, the only SSP giving a glimpse of that predicted viewability is Google, but, it is only given in 3 viewability buckets, and not the precise viewability value.

Second, the analysis of the Impact of actual Viewability on Ad Revenue was made based on the assumption that the predicted viewability computed by SSP/DSP is a historical average of past observations.


Impact of Predicted Viewability on  Ad Revenue

Measuring the impact of the predicted viewability bucket on Adex revenue allowed us to answer the following question: for the same ad unit, does Viewability have an impact on monetization?

For this, we took the probability of Viewability calculated by Adex, in a very fine granularity (ad unit, inventory type, size, device, custom dim, plus other unknown dimensions and no browser or country). This probability was provided in 3 buckets: <20%, 20-50%, and >50 %. We ran a one-month data correlation between the predicted viewability vs the RPM (Average revenue per thousand auctions) per GAM ad unit, device, and sites that were generating very low revenue:

Hence, we can conclude that moving to a “higher” viewability bucket implies an increase of ~30% in RPM. But, where does it come from?

The RPM uplift is mainly due to an increase in fill rate when moving to a higher Viewability bucket.

In relation to the CPM, we see a particular increase when moving to the highest bucket (>50%).

The results seem to indicate that improving your Viewability could generate a RPM uplift of ~30%, which is mainly attributed to an increase in fill rate for those ad units with a very low Viewability rate (<20%) and an increase in both, fill rate and CPM, when your viewability rate is higher than 20%.

To sum up, this analysis shows that managing to upgrade from one Adex bucket to the higher one can bring a +30% RPM uplift. Given the limitation of having 3 buckets and not the exact predicted viewability, we decided to run a second batch of tests in this analysis.


Impact of Actual Viewability on Ad Revenue

Measuring the impact of actual viewability on Adex: Prebid revenue, allowed us to correlate Viewability vs revenue at ad unit granularity.

For this, we took the Actual Viewability estimated from Active View Viewability provided by GAM on Adex. Then, we ran a 15-day data correlation between the actual viewability vs the RPM per website tag Pubstack x ad unit Pubstack x device:

Thus, there’s a positive correlation between Viewability and RPM which means that when Viewability rises, RPM is more prone to rise. Also, we wanted to check if there was any difference per device for Adex, and we found that there were different levels depending on the device:  

For Adex, we found that the impacts are at different levels. What about Prebid?

We can see an increase in RPM mainly when moving to 60% Viewability.

Therefore, we can conclude that it is complicated to specify how much the RPM will increase if Viewability increases in one point. However, we can see that there is a clear threshold of viewability per device:

  • Desktop: around 40% and 60% Viewability
  • Mobile: around 20% and 60Viewability

Moreover, we checked the link between Viewability and Bid Density (average number of partner responses per auction requests), to clarify if the uplift comes from more bids or from more valuable bids. We found that on most of the scopes, as the viewability increases, the bid density increases with a threshold around 50-60% viewability. 


Publishers should pay attention to Viewability. They should at the very least measure the Viewability Rate metric. Ideally, your Viewability Rate should be above 60% (threshold) on any ad unit.  There are some quick tips to increase your Viewability Rate such as implementing Smart Lazy Loading, as well as High-Quality Refresh, or In View Ad-Refresh. Pubstack can help you to implement these to get to your ideal Viewability threshold. 

Feel free to get in touch with us. We will be more than happy to help you to take your Viewability to the next level! You can always book a demo with us here.

Getting started is easy

You’re moments away from taking your advertising revenues to the next level!