Capture referral link and attach it to a banner?

So I was wondering if anyone has tried this. When you tap on an ad from AppBrain or LeadBolt, it usually first opens in a browser window first then it opens the google play store. I notice that the referral link is sometimes like ad.king.xxxx. Has anyone tried capturing that URL and then just attaching it to a banner or something in the app? That way whenever someone taps that banner and downloads the app, it is a direct conversion without having to see the appwall first.

Would they know if you did this?

They will know if you do that by looking at the HTTP referrer. Not sure if you will be able to spoof it though.

Why would I need to spoof the HTTP referrer? Each link has some sort of referral code in it. Unless that code is different for every user that clicks on the ad, I don’t see why it would matter. I thought they would know by looking at the adwall impressions and your CTR. The CTR would be extremely high in proportion to the impressions.

You can get an impression even when you don’t actually show the appwall by loading it into an invisible WebView, can’t you ?

You can use some random to click or not click action with 20% users and your CTR still will be high but natural

Why would this help your revenues - if those are based on eventual app install by the user ?

Or do you mean you would be cutting out some steps - and that would help conversion ratio slightly - i.e. clicking on “More Apps” on a screen takes directly to the Google Play page for that app.

Measuring installs

On a related note - most SDKs that do AppWalls - like AppBrain do NOT seem to require any extra permissions (i.e. they are not watching for installs - I think which requires a special permission in the AndroidManifest.xml ?) - so is it that they just wait a while and then just check if the app was installed before and if it is installed later etc. (which does not require extra permissions).

So one can see how SDK-based AppWalls can check if an app has been installed.

But how do non-SDK (for example HTML-based AppWalls) check if the AppWall eventually led to a real install ? Or are they REALLY only paying per click that takes directly to Google Play store (as that is all that a WebView can measure).

I guess so. But only if the if the adwall is HTML like for leadbolt. But AppBrain launches its own activity.

To answer your first question, it would be about cutting down the amount of steps. When people use an adwall, it is pretty apparent it is an ADwall. People don’t like ads so they may be put off by installing apps from a wall full of ads. So by removing the adwall completely and linking them directly to the app or game, it might increase conversions. The difference between an adwall and banner, in my case, relies on the fact that I make game recommendations in my app. That’s the purpose of my app. So I already have banners there for unadvertised games. So I can throw in ads for games in there and the user won’t know the difference.

And I don’t know the answer to your second question so I would love to hear how they track installs. I’m not even sure if AppBrain pays per install or not. All I see on my tracking page is eCPM.

AppBrain pays by installs. In fact GetJar has moved from not just installs but also that users open the app also - as that is the new “industry standard”.

Same situation with Tapjoy and the other AppWalls - most talk about “download and open this app” (so perhaps they are checking if app been opened also).

However these folks CAN check the install.

The problem is with banner ads and AppWalls that are HTML based - Admob/Leadbolt.

There I do not see how they can check for installs - so they are probably rewarding on just click. Which may explain why Leadbolt does not give an install figure for the AppWall.

AppBrain pays about $0.18 per install - Leadbolt AppWall and others probably pay more for installs - but I can only see the EPC (earnings per click) which is at $0.03 to $0.05.

From AppBrain, I can see that the actual clicks are about 3x as many as the installs. This suggests that per “click” the AppBrain revenue thus becomes 1/3 or $0.06 per click or so.

So in that way you could say that AppBrain and Leadbolt are paying much the same.

Except AppBrain actually (probably - or it CAN) check for the install of the app - while Leadbolt probably cannot because it is a WebView and thus is in the same boat as banner ads (and thus probably also more exposed to click-fraud).

Which leads to the next question - does the Leadbolt SDK-based AppWall pay MORE than the HTML based one ? (becuase with the SDK based one they should be able to reduce click-fraud).

Didn’t Leadbolt have a non-EULA version of the SDK for the display ads ? If so, that should be ok to use - EXCEPT Leadbolt probably does not have a prefetch API for the WebView (which you can do if you are doing via WebView yourself - as suggested in another thread).

Leadbolt should probably include a prefetch mechanism for it’s AppWall in the SDK - and then developers maybe keen to use that (also reduce click-fraud discounting of their revenue).

I emailed AppBrain regarding what non-banner revenue was and some other questions. Here was their answer.

The non-banner revenue is from the interstitial.
The payment is per install. The impressions are just shown for your information.

Eg yesterday your revenue was based on 57 installs. The installs come via the banners and interstitial combined. The tracking is accomplished without special permission via the SDK.

So I guess the impressions don’t mean anything. I have a hard time buying it because they also show eCPM. Why wouldn’t they show average PPI? That would make more sense.

Eg yesterday your revenue was based on 57 installs. The installs come via the banners and interstitial combined. The tracking is accomplished without special permission via the SDK.
So I guess the impressions don’t mean anything. I have a hard time buying it because they also show eCPM. Why wouldn’t they show average PPI? That would make more sense.

Thanks for that - it confirms what I was conjecturing.

The installs are then the total installs - the revenue is based on that - if you divide your revenue per day by the installs per day - you will get a fairly stable figure close to $0.18 per install (it maybe less these days).

For AppBrain:
Date Banner-impressions Banner-Revenue Banner-eCPM Clicks Installs Non-banner-revenue Total-revenue

The impression are for the banner ad only.

They do not give out the impressions for the AppWall.

Thus the full set of info is available for the banners. You can confirm that the eCPM is just a derivative number that is calculated from the banner-revenue/banner-impressions division.

The thing that is missing is the AppWall impressions per day.

The clicks may refer to just the banner clicks - but given it does not say Banner-clicks - it may be the clicks that you have when folks click on the app in the AppWall (whether shown as an AppWall or from the banner ad - since the banner ad ALSO shows the same AppWall).

So if that is true, then clicks would be user clicks on the app in the AppWall and the installs the total installs.

You can see that the clicks are about 3x the installs - which fits into this model. If it was clicks on the banner ad, it may have required more clicks to get one install (?).

David or others:

How is the performance of Tapjoy compared to AppBrain/Leadbolt AppWall ? If one uses Tapjoy just as an offerwall ?

If one uses Tapjoy as a virtual currency thing - I wonder if it will do better than GetJar (which has not done well so far - though how it is placed may make a difference also).

David mentioned in another thread earlier that Tapjoy was doing well but lately not done as well.