UA - 5 min read  - July 22, 2019

CPI, CPA, CPM or RevShare: Choosing the Right UA Pricing Model

Competition in the mobile market is heating up. Between March and May 2019, about 6,000 apps were added to both the AppStore and Google Play on a daily basis. But don’t get upset just yet.

The number of mobile users is growing as well, adding 100 million from Jan 2018 till Jan 2019. With such an enormous volume of apps and users, how can app marketers find the right users for their apps?

The answer is pretty obvious: paid user acquisition. But obvious doesn’t always mean simple. With increasing competition for users, paid UA becomes a cost sensitive topic. Mobile app businesses are constantly searching for the right balance between quantity and quality of users, as well as price. It’s easy to waste user acquisition budget if you don’t know your users and market.

Even though there is no silver bullet, there are some basic do’s and don’ts that can help you guarantee the success of your campaigns.

When it comes to paid UA, there are a number of pricing models to choose from. The most popular are CPM, CPI, CPA and Revenue Share. In this article I’ll give you our team’s insights on how to choose the right UA model for your mobile app business.

Let’s start with a general overview of the pricing models.

CPM (cost per mille)

CPM means the cost-per-thousand impressions. Generally the costs of such traffic are really low, and volumes are big. But for mobile this particular model is rarely used, due to bad performance of such traffic: low install and retention rates.

The top CPM use cases are branded campaigns aimed at awareness, not performance.

Such traffic is prone to fraud. Big volumes, a large number of traffic sources and no target actions make fraud protection an issue for both traffic providers and publishers.

CPI (cost per install)

CPI seems to be the most popular model. But does popularity make it universal and profitable? It really depends on your goals.

CPI works best for apps where ads are a sizable monetization source, e.g. hypercasual games. This pricing model provides lots of low-cost installs and targets users who can be monetized later via in-app ads.

This is also a suitable model for newly released apps. It helps gather a sustainable user base, test monetization models and game play, collect significant stats and track user behavior. Most importantly, this period can be used to collect statistics on installs and payments, which can help you figure out the right pricing model for your project.

The main metrics to pay attention to, particularly if you are using the CPI model, are retention rates, LTV and ARPPU (average revenue per paying user). They can be used to optimize your UA campaigns and detect the best performing traffic sources.

Fraud is also a big issue for the CPI model because installs are quite easy to fake and, until now, there haven’t been 100% effective anti-fraud solutions. If you are going to use CPI, get ready to put all the necessary fraud detection strategies in place.

CPA (cost per action)

Under this model publishers only pay for users that fulfill the target action. Such actions could be subscriptions, trial completions, purchases, etc. This traffic is usually more costly, but of a much higher quality.

The CPA model is best suited for games with in-app purchases or subscriptions as their main source of monetization, e.g. shopping apps or established mid-core games. These apps are focused on high quality traffic that fulfills those target actions and publishers are ready to pay for it accordingly.

The primary metrics for assessing the quality of traffic under the CPA model are retention rates and payment analytics. The latter is important if your target action is connected to payments. It’s vital to understand how much and how often users are paying in your app, so you can use this data to optimize your CPA costs.

Also, using the CPA model can leverage valuable ad LTV data. Ad LTV shows how much ad revenue the average user generates during his or her lifetime in the app. This metric is vital for apps that have a substantial amount of their revenue coming from ad monetization. It helps determine the amount of money a marketer is ready to spend on users monetized through ads.

As for fraud, the situation can be different for different app types. Target action that involves payment of any sort makes fraud quite meaningless. But in the case of trial completion, for example, fraud is still an issue and the incentive for fraud might be even bigger since the cost is higher.

RevShare or Revenue Share

Derived from the CPA model, the revenue share has a very specific targeted action — payment. Unlike CPA, where publishers pay a certain price for each target action completion, publishers using revshare pay a certain percentage of the money they received from users.

Revshare is the most risk-free model for publishers. Though it has its sensitive point — trust. For the model to work, publishers need to share info on all payments that users make in the app. So before switching to this model, it’s totally fine to run test campaigns and get to know your UA partner.

This model works extremely well for apps with big user bases and a good percentage of paying users. These publishers aim for whales and search for traffic sources that are able to target them. Whales are the top-spending users in an app, which makes them a tasty target.

The main metrics to pay attention to while considering a revenue share model are LTV and ROAS (return on ad spend). You need to know exactly how much users are expected to spend over their lifetime in your app and how much profit you can make from them. Then you are able to decide what percentage of that revenue you’re ready to pay for attracting profitable users.

Fraud makes absolutely no impact under this model, as all the risks are on your traffic partner’s side.

Choose your UA pricing model wisely

To sum this all up, here are the main points to pay attention to while choosing your UA pricing model:

  • What is the goal of your campaign?
  • How much traffic do you want to get?
  • How are you going to evaluate the quality of this traffic?
  • What protection do you have from fraud?
  • How are you going to analyze your users? What insights can you gain to optimize future campaigns?

Once you know the answers to all those questions, the choice becomes pretty obvious. There is one more important thing to remember though. Just as your app develops and grows throughout its existence, your UA goals can change as well. With changing internal and external conditions, other pricing models can become more profitable and meaningful than what you currently use.

I’m going to talk more on this topic in the next article. Make sure not to miss it.


Our AppGrowth solution specializes in acquiring high quality users not only through the world’s largest RTB exchanges, but also via our exclusive inventory of 33,000+ app with over 500 million MAU. AppGrowth performs best under the CPA and RevShare models, but the CPI model is also available as an option. Find out more about its other benefits here.

If you are looking for sophisticated user insights, check out our analytics solution DataCore aimed at providing deeper insights into your audience and marketing channels.


Tanya Kovalevskaya
Product Marketing Manager