One of the hidden gems in the SDKless ad quality solution we recently announced is the churn screen. This screen allows you to see what advertisers are causing more churn in apps that are part of the SOOMLA community.
This post is going to be about a set of ads limited to a single advertiser and a single network causing 15% churn to publishers who these ads run on their apps. This means that out of every 100 impressions that users saw, about 40-50 clicked on the ads and 15 didn’t come back to the publisher app for at least 7 days. This number is very high – if all ads were taking the same toll the game would quickly run out of users.
Who is the advertiser and the network?
So while the identity of this advertiser and the network is available in AdIntel, we don’t feel it would be right to point the finger in this blog post.
We can say however this is is a matching app with level based progression.
Analysis of the advertised app causing churn
Overall we can say that this advertiser is highly focused on Rewarded formats and that Playables are giving him the highest CTR
His top 4 networks are:
His US CPI is $2.47 while IPM is 5.14
Analysis of the network causing the churn
The network causing the churn is one of the top 10 and most of it’s activity is in US and Korea.
The advertiser of the app in discussion is the 13th customer for the network when sorting by budget size globally and the advertised app is #34 for US-Android.
The network in discussion is serving almost exclusively playable formats as shown on the left side chart and those are split into interstitial-playable and rewarded-playable (including interactive end card formats). The category focus for this network is Puzzle and Casual.
Analysis of ads by the advertiser and network
We use the “Ad UX Drill Down” section to further analyze the ads by this advertiser and network combination. There are 42 ads by this combo and we inspected the top 5 by budget.
The 1st thing that grabs our attention here is that CTRs are quite high. Some of the ads generate over 50% CTR while others are not far behind. Keeping in mind that this is a matching app and it’s quite hard to get users excited about this gameplay anymore – 40-50% CTRs values are very high.
The bounce rate is not shown in this table (will be added soon) but we can see that the QCTR is less than half of the CTR so bounce rates are around 60%. In other words, 60% of the users regretted their click and came back right away. About 80% of the users that didn’t came back right away, never came back or at least didn’t come back for 7 days which is the reason why we are looking at this set of ads in the first place.
Looking at the ads themselves we can see that they all have 2 parts:
- Video of about 30 seconds
- Playable experience that lasts at least 15 seconds
When we are looking at the heatmaps we can see that many of the clicks that resulted in the opening of the app store were actually attempts to close the ad that the ad refused to acknowledge as close attempts.
We can also look at two other aspects – this is from one of the top ads –
The top visual shows us that when users do click on the ad, they are not immediately taken to the app store, instead, the ad would wait and trigger the click much later. This means that it’s likely the user didn’t expect his click to trigger the app store opening. Especially considering all these ads have playable end cards it is likely that users where simply engaging with the playable component.
The bottom visual in the screenshot below shows that the ads are pretty much impossible to close before 45 seconds have passed and many of the users gets trapped in them for over 60 seconds.
So we started with a set of ads we found by a specific advertiser and network that were causing high churn for publishers. When drilling down on the these ads we found some alarming patterns that could explain user retaliation towards the publisher. We didn’t show this in the post but there are other posts we published showing that these experiences actually reward the advertiser side so they are not likely to go away on their own.
Ads causing high churn are a real threat
Over the course of 2 weeks our system detected over 900 high churn scenarios. These are specific sets of ads for one advertiser-app, by one ad network that are damaging for publishers. In each of these scenarios, at least 5% of the impressions led to the user churning. There are 2 ways for publisher to overcome this problem:
- Implement an SDK based ad quality solution – in 3-6 months they should be in a position to eliminate the problem from their apps
- Use the SDKless Ad quality solution as part of AdIntel – in a few days you can ask the networks to block the problem ads from your app
We can help with either option or both.