[gyardley]

I should probably write a blog post on this too.

Essentially you buy inventory from a specialized ad network, and they charge you by the install, using UDIDs to match the users who click on the ads with the users who installed the app. That way you get charged only for the users they actually sent you.

Most application ads convert poorly, so the companies doing the most volume in this space run 'incentivized' ads, where the user gets some virtual currency for installing the app.

For example, if they're playing 'Tap Widget', which normally charges for widgets, the user can get a widget for free by installing an application.

The user wants the widget, so they install the advertiser's application. The ad network sees the install and tells Tap Widget to give the user his widget. The advertiser gets a new user, so the ad network gets paid. The ad network takes their cut and passes the rest of the money on to Tap Widget. This works out pretty well for all parties.

I own a chunk of Flurry, one of the companies that specializes in this. There are others.

[nobody_nowhere]

Interesting! That's a pretty clever way to goose your CPA performance.

[gyardley]

Thanks! I can't take credit for it, though - incentivizing CPA offers was probably first done by Gratis Network a decade ago. That was for actual physical goods - complete five offers, refer three friends who do the same, get a PlayStation or an iPod.

Later, the model was adapted for Facebook games, with the offers remaining the same (insurance quotes, etc.) but virtual currency used as the bait. On iOS, the focus is on application installs, but it's still pretty similar to the Facebook model.

I can only recommend the model under very particular circumstances, since what you gain in conversion rate you lose in traffic quality. If users are shown a list of ten different offers, they'll pick the offer that sounds most appealing, but at the end of the day they really just want the incentive. Companies like Gratis were constantly churning through their advertisers, as each advertiser figured out that most of the traffic they bought couldn't care less about their products.

On iOS, the advertiser churn doesn't happen, because the installs are just a proxy for front-page App Store placement - and at the moment, anyway, App Store placement doesn't have anything to do with application usage. Because of this, the advertisers get what they want, consider any application usage they get a bonus, and the incentivized model keeps working. If Apple were to change the 'top free' list to take into account time spent in app, the model would have to be changed.

[nobody_nowhere]

Apple probably doesn't care much about how much you use the apps, as long as you come back to the store. Even for the free apps, they're probably thinking primarily about store engagement rather than app engagement.

As an apple shareholder, I like revenue optimization, but I can imagine it's pretty frustrating if you have a horse in the app race. Or out of a general sense of fairness :)