Historically, there’s never been an advertising program that actually guarantees results. Things change though, and today different advertising networks are offering CPI or Cost Per Install advertising, possibly subject to some limitations. Let’s take a look at why and when you would want to go with a CPI model vs. a Cost per Click or Cost per 1,000 views advertising model.
In traditional advertising, the person or company purchasing the advertising assumed the responsibility and risk for the advertising performance. It was up to the people buying the advertising (in whatever form – billboard, television, radio, direct mail, banner ads, etc.) to produce and test their ads to make sure it gave them the return they expected. Cost per Install changes that, the advertising network (venue) takes on the risk. It is one method, not the only method and it requires investigation on a per app and per marketing campaign basis to determine which model is likely to be most profitable for the developer. One shoe will not fit everyone equally.
Cost per Install is an attractive model when you absolutely know that the profit for each download or the collective average of them, will exceed your CPI buy rate and overall break even point. Your break even point considers the time/wages invested in developing the app, your office overhead (like electricity, internet service), and so forth.
If you are making $4.00 on every app and it only costs you $1-2.00, you could kick back and never have to worry that your app would stop being profitable. Even if you are making only $.01-02 per install, you will always be able to buy more advertising. Premium paid apps are likely to fit this model best. After someone installs and pays for your app, it is mostly inconsequential whether they ever use it again. If, however, you depend upon in-app advertising, in-app currency or upgrades, then obviously your profit margins are not guaranteed. CPI could be a quick way to the poor house for you.
With all other forms of advertising, you take on the risk that your advertising will perform. The way to mitigate that risk is to test your advertising with an audience. This involves setting up at least two versions of each advertisement you want to run to compare response rates on each. This is AB and multivariate testing using small sample groups (500 – 1000, sometimes more) to see which they like most, click on more, etc.
If you want the ad network to assume the risk, you pay higher. If you take the risk, you pay lower. This is an ages old, global paradigm. When it comes down to it, the profitability of Cost per Install advertising depends entirely upon how much you absolutely know you will make. However, if you have done your marketing and sales research, all other advertising models can generate higher returns for you.
Let’s use $1.00 for CPI and CPM as a fictitious example. To make par, CPM needs to generate a .1% install rate, one tenth of one percent. With CPI, you absolutely that you will pay $1 for 1 install. With CPM, it is possible that you could get less than a .1% install rate. If you’ve done your marketing and sales groundwork (research and testing), you could get a full 1% install rate, or higher. That’s a 10 to 1 ratio.
That’s something to seriously think about. Learning a bit of marketing, doing some analysis, doing some testing could increase your profitability by 1000%.