February 7, 2013
Guest post by Android app developer, Jason Hudgins.
Every mobile advertising ‘best practices’ guide for publishers harps on the fact that you should strive to increase your CTR, via a variety of means, in order to increase your revenue. Since virtually all mobile advertisers use a pay-per-click model, it’s a rather straight-forward assertion that the more ads in your app that get clicked, the more money you’ll make.
Beyond this, however, there is often an implied message, that the higher your CTR, the better ads the network will send your way (in terms of earnings per click, or eCPC). Perhaps this is true for some, but I certainly haven’t found this to be the case. eCPC is likely influenced by a variety of factors, most strongly by seasonality. Specifically to note, is that in my experience a higher CTR hasn’t shown to result in a higher eCPC.
So assuming that they are not correlated, what other ways can we optimize our CTR? The reality is that one of the few things as a publisher we have control of is our ad-refresh rate. There is indeed a very strong correlation between how long a banner ad is displayed and your CTR, however that alone doesn’t tell the whole story.
CTR is expressed as a percentage of ad impressions. So if you were to show 1000 impressions over a 60 minute period and only 1 ad was clicked your CTR would be 0.10%. Alternatively, if you were to only show 1 impression during 60 minutes (humor me here) and that 1 impression resulted in a click, your CTR would be a perfect 100%. In either case there has only been a single click and you’ve made roughly the same amount of money. In short, CTR is an interesting stat, but it’s our overall revenue we really care about.
So here goes—my logic (based on the performance of my current app) to earn higher revenue:
You need your users to click on more ads more often, and if we focus only on ad-refresh times we want to find the sweet spot. The metric you are looking for then is clicks per session. What you want to do is take the number of clicks during a given time period and divide that by the number of user sessions during the same time period:
( total clicks / # of sessions ) * 100 = percentage of sessions resulting in a click (higher is better).
As a simple concrete example, say if during a two-week stretch I had 250 clicks from 10,000 user sessions:
(250 clicks / 10,000 sessions ) * 100 = 2.5%
That just means that 2.5% of all user sessions, result in a click where you earn some money.
So to kick-off the new year with a new view (and new revenue), my suggestion is that you apply this formula over time, while testing it against a variety of ad-refresh rates. All other things being equal, eventually you should find the optimum refresh rate that results in the highest percentage of sessions resulting in a click.
Next month, I’ll be sharing data from my app that directly correlates to this metric, so stay tuned!