Greystripe and Flurry announced cost per install models today for mobile application installation. They are solving a distribution problem for iPhone developers, one the iTunes store caused: little shelf space.

The equation governing ad spend for a developer is:

cost of customer acquisition < average revenue per user

For applications monetizing with paid downloads at an average price of $1.29, it's hard to justify a CPI model charging $0.99 per install. After Apple's cut, you're left with $0.14 of revenue per user.  It is better to give a user a dollar to buy your app, give $0.30 to Apple and keep the $0.70 for yourself. Flurry's model specifically cuts the CPI rate to $0.25, a step in the right direction for most developers. 

With that in mind, it will be interesting to watch whether CPI on the mobile evolves like Slide and RockYou on Facebook or like Zynga who spends millions per month on CPC and CPM advertising on Facebook. In the first model, venture dollars were being spent by one app developer, paid to another developer and a revenue share was extracted by an ad network. Unless apps can generate large revenues from users, eventually the venture dollars run out. In the second, average revenue per user massively exceeds the customer acquisition cost so large scale paid marketing is profitable and spend increases.

Without a doubt, CPI is the right model for developers seeking users – it aligns the incentives of ad networks with the developer. The same is true on the web. But by and large cost per acquisition (the equivalent on the web), hasn't been successful for Google or other ad/search networks. Rather it has found a home in lead generation and affiliate advertising. 

It's unclear what will motivate large scale adoption of CPI on the web and for mobile. Lower prices may be the answer.