So you are saying in one case you were giving more impressions than the other - but CPM being still the same (by my theory it should have decreased as user fatigue not able to keep up with the increased viewing) - that means impressions x CPM = revenue - was greater.
Interesting.
I think my long-winded comment above was more geared to analysis of banner refresh - I have set the admob refresh rate from 60 seconds to 120 seconds based on that sense.
Now you say that you are getting more revenue with more impressions - which is very interesting.
But I think your case is slightly different from banner refresh.
But it is an interesting statistic nonetheless.
quote:
for 1 client I implemented it on launch and on exit.
for another client I implemented it on launch, on exit, and during natural breaks in the app.
in both cases the CPM is the same.
In one case you are getting nearly 2 showings of the Airpush Smartwall in one app.
In the other app you are getting 2 showings (beginning and end) PLUS some more (I am guessing 2-4 times for between level type showings of Airpush Smartwall).
It is a different situation - but suggestive that users are likely to click on the Airpush Smartwall regardless of how often it is shown (though it would be not more than once per minute or once per two minute I am assuming). So it is not the same situation as banner presentation - plus appwalls require user interaction and cannot be ignored.
This very fact should make interstitial ads more valuable than banner ads. Plus they are guaranteeing attention at least once every 2 minutes (something which is not assured even with the banner ads refresh of 60 seconds or 120 seconds - unless the banner graphic draws the user’s attention).
Let me analyse the various rationales why appwall presentation could give this linear response (just guessing here).
If the CTR is almost entirely based on “oops I mistakenly clicked” - then the “clicked by mistake” CTR should remain constant (thus the total click-throughs should scale with the number of impressions).
Were you getting a very low CTR (which would be suggestive that you are operating in this regime ?).
If on the other hand users are operating in a regime where they are averse to the showing of the appwall and generally avoid clicking - then more frequent presentation may make them feel they should click more (and that the appwall is being presented more precisely because they have not clicked it before).
In any case that is a very interesting statistic.
Did you get any user feedback that complained of too many ads - or it was all acceptable.
I am trying to do something similar with greystripe interstitials - however while appwalls pay for click-through (or install of the app on offer), on the greystripe they say it is by impression. This should be similar for video ads like adcolony I am guessing.
Except since I had a voice changing app I had no natural breaks where I could present such stuff. The user interface is pretty simple.
Eventually I broke up the user behavior as comprising essentially a record/play cycle - and settled on a lazy display of ad strategy - where ads could only occur after a play ended. That is an ad would be shown only when the user became idle after having completed a mini-use-case (I am giving it 4 seconds of idle after a play ends in order to show an ad). Then there is an additional cap (to prevent overpresentation of the ad) that prevents such ad presentation too often (can’t be less that 2 minutes between ads).
I will see how this works - although I will have no comparison data (for example if I had a version of the app where I just showed the ad at start).
In my practice with the app I am getting a sense that this strategy IS working as the ads do not feel intrusive. The strategy works also with the greystripe ads (at least with the test ads) because the user cannot inadvertently click on the ad but a click presents a skip and a click option - this way the user has a way to skip without inadvertently clicking. Also this type of ad is impression based - so it should technically (assuming fill rate is there) give a predictable performance (since there is not another intangible of user-inclination-to-click-through as with an appwall).
So I will see if this works for my particular app use-case. As most-importantly the ads NEVER show up when the user is frenetically using the app, record/play or when adjusting some setting. The ad seems to magically appear when they pause to contemplate something - or when they look up at someone that “hey, you try it”.
In games where they have levels - the between level gives them a natural break - so the model I use may not be appropriate for them.
But it may be appropriate for games/apps where there are no natural breaks in the use.
I would welcome comments on this way of presentation of ads.