CPI bidding is the most common and is safer for the advertiser since it is guaranteed to hit the entered cost per install (CPI) goals. Since AppLovin risks potentially serving a campaign with broken tracking, CPI campaigns may have to be validated by an admin before the campaigns launch. CPI campaigns may take longer to scale because the optimizer will hold the campaign in the learning stage longer to gain more significance on performance. Also, CPI campaigns may overshoot budgets since there is a longer delay between when the ad request happens and when the install comes in. For example, if installs tend to happen three hours after an ad request, the ad server will stop serving when the campaign hits the budget but three more hours of installs will roll in after that.
Dynamic CPM (dCPM) is a bidding method in which advertisers are charged for impressions, but the optimizer will adjust the CPM up and down in order to back out to a CPI goal. CPM bidding comes out of the optimizer’s learning stage fastest, which allows campaigns to scale the quickest, however the initial days of the learning stage may result in inflated install costs. CPM bidding is typically leveraged by advertisers with larger budgets who seek immediate scale.
Dynamic CPC (dCPC) is a bidding method in which advertisers are charged for clicks but the optimizer will adjust the CPC up and down in order to back out to a CPI goal. CPC bidding has the added advantage of coming out of the optimizer’s learning stage faster than CPI bidding (though slower than dCPM bidding), which will cause it to scale more quickly. CPC bidding is recommended for advertisers with larger budgets and a higher tolerance for learning costs.
Campaigns of all bid types will have some costs associated with learning. For CPI campaigns, the cost is for the most part borne by AppLovin, but this will cause the optimizer to be more conservative initially. For dCPC campaigns, it will take some budget before the optimizer hones in on your goals. The amount of budget it will take depends on targeting, conversion rate, CPI goal, and budgets.
The more broadly the campaign is targeting, the more places it has to learn.
The better the conversion rate, the more quickly the optimizer will learn. You can increase conversion rate by taking actions like uploading a suppression list, wifi-targeting larger apps, and having strong ads.
This goes hand-in-hand with conversion rates. If the CPI goal is very obtainable, this will decrease the time it takes to learn. If the goal is too low, the optimizer will take a while (or will never be able) to hit your goals.
Start with the highest bid you are comfortable with. This will give your offer the best chance at success. Dropping the CPI later if the offer converts well is better than increasing the CPI later if the offer doesn’t convert well, in terms of how the optimizer serves the offer.
Starting too low can also lead to a situation where the offer spends a lot of money but gets shut off. For example, if the CPI goal is unachievably low, the bidder will start shutting off sources after it has spent learning budget on those sources, which can lead to a situation in which an offer spent money with no results.