How to evaluate the effectiveness of advertising in the application: ARPDAU, seasonality and a few secrets

More and more developers today monetize applications with advertising. How to correctly assess how effective the advertising service that you have chosen? We propose to use ARPDAU, and not only in the classic sense of this indicator.



ARPDAU (Average Revenue Per Daily Active User or average revenue per daily active user) is an important metric that allows not only counting revenue, but also evaluating the effectiveness of any advertising mediation solution. Here is how it is calculated:

ARPDAU = Total Revenue / DAU (Daily Active Users)

But just one total ARPDAU is not enough to correctly evaluate the effectiveness of your advertising partner. To make decisions about mediation or monetization strategies, ARPDAU should be evaluated in terms of several parameters. Consider the examples.



Pay attention to the distribution by country

')

Imagine this situation: you have noticed a clear decline in revenue or eCPM, but you cannot discover the source of the problem. At the same time, there were no visible changes on the part of the mediation service and the monetization strategy you did not change. Check out the breakdown by country and see how the Tier 1, Tier 2 and Tier 3 indicators change over the months. For an objective result, it is important to consider specific countries - as a rule, this is your top in traffic and / or revenue. For example, the United States - the region of Tier 1 - can bring the application 30-40% of revenue.



Suppose your top revenue is the United States, and the proportion of the regions has changed - for example, American traffic has decreased, and a less profitable region (for example, Tier 3 - India or Russia) - has grown. In this case, revenue significantly fall, because US traffic is much more expensive than Tier 3.



Evaluate the effectiveness of types of advertising



Want to know the performance types of advertising or replace one format with another? The scheme is the same: watch for ARPDAU changes for different advertising formats and the dependence of this indicator on the number of users. The calculation of ARPDAU in the case of advertising formats is an additional way to analyze their effectiveness. For example, you can compare the indicators of native advertising and banners, or find out what works best - full-screen advertising every three minutes between levels or a permanent lower banner on the screen.





Different types of advertising in the game Fleet Battle by Smuttlewerk



Do not forget: the change in performance of different advertising formats and, as a result, ARPDAU, also affects the Display Rate. Display Rate - a metric that displays the ratio of the displayed advertising to the loaded. It is a poor performance and a decrease in ARPDAU that can cause a low Display Rate. Read more about Display Rate in the article at the link .



Compare by application version



For more detailed analytics, you can compare ARPDAU for different versions of the application. For example, using Google Analytics, you can calculate the ARPDAU version with one monetization option and compare it with the ARPDAU version with another option or with a version in which advertising mediation was not used at all.



Android applications have a separate option: you can conduct an A / B test of two types of mediation using Google Staged Rollout - and in the process, calculate and compare ARPDAU.



! Many people underestimate the Google Staged Rollout tool, but it can be very useful. This tool allows you to test not only different types of mediation, but also safely upload application updates. For example, with its help, you can download a new version of apk and “roll out” it by 1% to 50% of the audience with random distribution of traffic. Also, the first few days after the release of the update, you can monitor possible errors - and in case of their absence, “roll out” the application by 50%, and then to 100% of the audience.





Source: Jana Technology



Consider seasonality indicators



Also, when testing mediation, one should take into account seasonal changes in indicators. To avoid errors, exclude the months in which the indicators change in steps. Usually the most profitable month is December, and the lowest revenue and eCPM are expected by developers in January. For example, according to Appodeal, from December 2016 to January 2017, eCPM for banner advertising decreased by 46%, for Rewarded Video - by 9%, for full-screen video - by 18%. In order for the evaluation to remain objective and not affected by holiday and post-holiday jumps, these months should not be taken into account in the process. Read more about how to make money in the holiday season, read this article .





Source: Appodeal data



For iOS version of the application, we recommend performing the test quarterly, taking into account seasonality. The more data - the better: take the data for at least 3-6, and better - 6-12 months. If possible, feel free to collect DAU (daily active users) by top countries (by traffic or by revenue) for this period of time and read ARPDAU.



Remember that after switching to a new advertising mediation service, you should wait a few months to get a set of data on the average ARPDAU by country already with a new advertising provider.