Ad Refresh and Its Impact on CPM

We have talked about ad refresh extensively in our blog and as you here, we assume you are familiar with the concept. While ad refresh has been here for the last few years, one of the most pressing questions – ‘what will be the impact on CPM?’ remains. 

On one hand, advertisers want to ensure that the ads aren’t getting refreshed too often, especially, without considering viewability. On the other, publishers want to refresh to maximize the impressions and the overall revenue. In general, everyone in the industry accepts ‘ad refresh’ provided certain conditions like viewability, time-in-view are met. 

But there’s no standard from any of the industry consortiums. And, that’s why it’s important to refresh based on viewability and user activity on the page. If you do it right, your inventory value isn’t going to be depreciated. 

“An ad that appears later in the refresh sequence may be less prone to the accidental clicks that account for most of the high CTRs of the first impression, but it is not necessarily less effective at delivering brand lift, recall or conversions.” 

AdExchanger

At Automatad, we offer ‘Active Exposure Time (AXT)’, an advanced ad refreshing technology that refreshes ads based on ad viewability and user activity on the page. To put it simply, we wouldn’t refresh:

– out-of-view ads,

– Ads that aren’t seen by an active user on a page.

– Ads that stayed in-view to a user who isn’t active on the page.

Ads that are in-view to an active user will get refreshed. If there’s an ad that isn’t in-view, we wouldn’t refresh it even after 60 seconds. After all, it is important to align the incentives of both publishers and advertisers. 

But the questions require data as an answer. In this post, we’ll take a look at how AXT impacted several metrics including eCPM, ad viewability, and total revenue. Shall we begin? 

First Impression and AXT Impressions:

‘First Impression’ refers to the first ad of an ad unit that’s been rendered to the user. It’s the first in the sequence of ad impressions. An ad unit can render multiple ad creatives one after another. How long an ad creative stays on the page/ad unit will be decided by the triggers set by the publishers or the partner. 

For instance, you can use a time trigger and refresh ads every 30 seconds using Google Ad Manager. Else, you can use a combination of triggers (as we explained above – AXT) to refresh ads. As we’ve used AXT to refresh ads across our partner sites, we’ll call the consecutive impressions (second, third, etc.) as ‘AXT Impressions’.

Limitations:

a. We can’t attribute the increasing eCPM and revenue to AXT and we aren’t. The point we are making is that the eCPM/revenue increased in spite of it, not because of it. 

b. eCPM typically increases during Q4 and that’s exactly why we analyzed data during the same period. We wanted to see how eCPM varies (for the First Impressions) in the presence of ad refresh. Has the growth in revenue/eCPM slowed down? No, it didn’t. We saw a 45% jump in eCPM.

c. If you notice the graphs, you wouldn’t see a smooth linear growth in eCPM or revenue. There’ll be quick ups and downs because of the noises in the datasets. We haven’t employed any data cleaning methods to enhance the analysis.

First Impression Data

We analyzed over 1.015 billion First Impressions (FI) across ~150 web properties to understand the distribution of eCPM, ad revenue, and ad viewability during the experiment period (Q4 2019). 

Audience Geo Split-up: India – 50% | U.S – 15% | U.K – 5% | UAE – 4% | Canada – 3% | Australia – 3% | Rest – 20% 

a. FI eCPM:

From the graph, we can see that the eCPM has increased over time and it isn’t getting impacted negatively because of our ad refreshing product

FI - eCPM Distribution Across the Network Over Q4 2019b. FI Revenue:

As expected, ad revenue increased over the 3-month period and it continued until the last week of December. FI - Ad Revenue Distribution Across the Network Over Q4 2019c. FI Ad Viewability:

Ad viewability neither increased nor decreased throughout the time period. It makes sense as the viewability of the First Impression isn’t influenced because of refresh. Think of it this way: AXT will have no control over the First Impression as it starts to consider viewability and user activity from the second impression onwards. We’ll talk about the impact on viewability later in the post.  

FI - Ad Viewability Distribution Across the Network Over Q4 2019 (1)AXT Data

Similarly, we analyzed ~ 600 million ‘AXT Impressions’ across the same web properties during the same time period. 

Audience Geo Split-up: India – 49% | US – 15% | UAE – 5% | UK – 4% | Aus – 3% | Canada – 3%

When it comes to AXT, the story is quite similar. eCPM and revenue increased (note that this inventory is being refreshed) and ad viewability experienced a slight improvement as well.

a. AXT eCPM:

Interestingly, eCPM of AXT impressions (refreshed ad impressions) increased over time — the same way it did for the First Impressions.

AXT - eCPM Distribution Across the Network Over Q4 2019b. AXT Revenue:

As eCPM increased, the ad revenue followed the suit. You can see the ad revenue of AXT Impressions started raising at the beginning and reached the peak value near the end of Q4.

AXT - Ad Revenue Distribution Across the Network Over Q4 2019c. AXT Ad Viewability:

Ad viewability of AXT Impressions slightly increased, but there’s more we want to convey in terms of viewability.

AXT - Ad Viewability Distribution Across the Network Over Q4 2019It is evident that when you refresh ads based on viewability and user activity, 

  1. eCPM and revenue of First Impressions aren’t going to be impacted. They tend to perform as they usually do. 
  2. eCPM, revenue, and ad viewability will improve for the refreshed inventory. While the degree of improvement depends on several factors including seasonality, demand, the market value of the AXT Impressions wouldn’t decrease over time. 

Impact of AXT on the First Impressions:

It may not be clear from the separate charts, but when you overlap data from FI with AXT, we can derive a few more interesting inferences. 

a. eCPM:

While the eCPM of refreshed impressions (AXT) isn’t the same as that of First Impressions (FI), both follow the same trajectory.

FI Vs AXT eCPMb. Ad Revenue:

Ad revenues from both impressions varied alike. As the volume of served impressions is different, the revenue of First Impressions (FI) is higher than the revenue of AXT impressions.

FI Vs AXT Ad Revenuec. Ad Viewability: 

This is perhaps, the most significant change happened because of AXT — Ad viewability jumped up to 50%. As AXT refreshes ads only when 1) a user stays active on the page for a certain period of time (say 25 seconds) and 2) the ad is viewable as per the MRC viewability standard, the ad viewability of the AXT impressions shot up immediately.

In fact, it continued to increase over the time period.

FI Vs AXT Ad Viewability

Conclusion

So, did the eCPM or revenue of First Impressions deviate from its usual trajectory? No.

Should publishers consider refreshing ads? Yes, they can refresh ads, but ad viewability and user activity on the page should be considered.

Will refreshing ads increase the eCPM of the First Impressions? No, it didn’t increase the eCPM of the First Impressions.

Can I increase ad viewability with AXT? Yes, you can. We’ve seen a significant increase in ad viewability across the network.

Is the eCPM of the AXT Impressions is same as that of the First Impressions? No.

Not but not least, ‘ad refresh’ can increase your overall revenue (completely incremental) and helps you to maximize revenue per page.

Leave a Reply

Your email address will not be published. Required fields are marked *