AppsFlyer is a mobile app analytics and attribution platform that helps you monitor, measure, and optimize your marketing campaigns. Appsflyer was founded in 2011 as one of the first app analytics platforms. The platform integrates with multiple ad networks, enabling accurate and advanced measurement. The tool itself focuses on measuring the lifetime value of the customer (LTV) , which is the most profound difference compared to web campaigns. Appsflyer also develops technology for measuring iOS users using SCAN 4.0 and the so-called Single Source of Truth .
Contents
Appsflyer technology itself focuses on monitoring the mobile app user over time, with particular emphasis on integration with advertising platforms, using SKAN to measure iOS and access to raw data. The most important functionalities of Appsflyer are:
The first, natural alternative to Appsflyer are native consoles, i.e. App Store Connect and Google Play Console . These internal analytical systems allow you to monitor the most basic metrics such as:
Importantly, these consoles do not require any SDK packages to be implemented in the mobile application. They are simply available after the application is published in the app store. They often operate only with estimates, do not contain any information about what happens after the application is installed, there is no access to custom reports and no support. Therefore, they perform rather illustrative than analytical functions.
The second group consists of mobile app measurement partners (MMP) . Each system has slightly different definitions, its own user monitoring specifics and its own integrations. There is no way to unify the data in each system, so there will always be discrepancies between them! Google Ads has created a list of recommended MMP partners :
Additionally, Google Analytics 4 (Firebase) is one of the alternatives for mobile app tracking, essentially free in the analytics part.
A bit earlier, Facebook Analytics was also a very interesting alternative, but it is no longer supported. Only the so-called event manager within FB/Meta remains. It will still allow you to analyze:
Regardless of whether you use native consoles or Google Analytics 4, sooner or later you should consider supplementing your analytics with an MMP partner. This will primarily allow you to track multiple traffic sources in one place (a large number of direct integrations), extend iOS user monitoring, and deeper remarketing analysis.
Appsflyer charges 0.0x$ for each non-organic event tracked in the system. Depending on the package, it can range from $0.05 to $0.08. The packages include all the basic functionalities except for Xpend for monitoring media costs, people-attribution mechanisms or access to raw data (from paid media).
Appsflyer will continue to provide:
The first step will be to register an administrator and add user accounts . Within Appflyer, you register applications, each for a separate operating system.
The second step is to perform the basic setup of your Appsflyer account:
Step | Description |
---|---|
Adding an app to your Appsflyer account | You can enter the application adding wizard in the Appsflyer main panel and go through the next steps |
Set a time zone and currency that suits your target audience and app needs | Once costs and revenues have been recorded, you will not be able to edit currency-related data. |
Attribution window setting | The re-attribution window is a period of time starting from the first install date during which re-installs from the same device are not considered new installs. The default is 90 days, which can be adjusted as needed. You may even want to consider extending this window to better focus on the new user. |
Once you are ready with the basic settings, it's time to implement the Tracking SDK.
Step | Description |
---|---|
Select the SDK features you want your app to use | This is an important area for advanced use cases, for example: will you use deep links? how will you implement privacy settings in iOS 14.5+? do you need to measure uninstalls? |
Select events to track | Think about what you want to measure in your mobile app. If you don't have a plan, you can use the recommended event generator from Appfslyer. The tool will suggest events tailored to the app category. |
Choose your attribution method and deep link | Typically this will be the Push API, but there are several other features available . |
Code implementation | This is a job for the development team. |
Testing | You can test SDK integration before or after your app goes live. You can use your app here for debugging – remember to submit your device for testing beforehand! |
iOS App Release | iOS app developers must provide information about their privacy practices when submitting their apps for review: what data they collect, for what purposes, and how it is secured. |
The above steps basically exhaust the configuration and developer implementation stage. The rest does not require coding or structural changes in the application itself. However, it is worth setting the elements related to attribution at this stage of understanding:
Step | Description |
---|---|
Integrate with advertising systems | Connect your ad accounts in systems like Google Ads, Meta Ads, Tik Tok ads to track traffic, user acquisitions and events |
Set up OneLink to attribute your own channels | To properly track traffic to your app from your own channels, such as your website, social media channels, or others, you need to generate the appropriate links using the OneLink tool. |
Set up deeplinks | You can set up deep links that will redirect app users to a specific screen, or so-called deferred deep links that will redirect new users first to the app store where they can download the app, and then to a specific screen in the app (it doesn't necessarily have to be the home page, you can redirect the user to any screen!) |
Set attribution windows | The majority of users, a little over 97%, launch an app within 7 days of interacting with an ad. The remaining 2% launch between 7 and 30 days (see the chart below). However, you can adjust how long after an ad interaction you want to measure attribution. |
Launch remarketing campaigns | If you will be running remarketing campaigns, first configure the conversion window in the application settings. |
Before we get to Appsflyer itself, it’s worth asking yourself what you want to measure, and how the selected metrics will affect the assessment or implementation of marketing goals. At this point, it’s a good idea to build an analytical plan.
An example of such a plan is dividing metrics into groups that examine different types of activity in the application. Some of them will be more useful to development teams (errors, load times), others will be used to examine user retention, and still others will focus on revenue or return on investment.
Each of the app marketing players, Google Play, App Store, Apple, Google Ads, Meta Ads and X Ads, as well as mobile campaign companies, have their own attribution models. Each of them counts installs and events slightly differently. And the point here is not to achieve full compliance at all costs (which is impossible), but to understand the definitions and be able to assess trends and draw conclusions.
In AppsFlyer, event attribution can be divided into two categories:
If there is more than one valid engagement during a new install, AppsFlyer will prioritize clicks over impressions and deterministic over probabilistic methods. In other words, where possible, Appsflyer will use full and verified data such as Device ID , and in other cases will rely on data from a provider such as SKAN 4.0 . It all depends on the possibilities offered by the specific situation: operating system, privacy settings, user tracking acceptance, and obtained data from the ad space provider.
The key to understanding Appsflyer reports is to differentiate the reporting methodology. AppsFlyer data is lifetime value (LTV) or specific activity (events) . LTV data is for installs that occurred within a specified date range. Activity data is all events that occurred within a specified date range. As you can see, in the case of LTV, both the time of installation and the time of the event are important, while in the case of activity, only the time of the event is important. Appsflyer installs are neither an event nor LTV.
ACTIVITY
Activity is the events performed by app users within a specific date range. An analogy to web campaigns can be used here. You simply select a time period and verify sales value, costs, return on investment, engagement, etc. You could say it is a chronological analysis and is useful for short-term verification of marketing effectiveness.
LIFE TIME VALUE
LTV data is events that occur throughout the lifecycle of a user who installed your app within a specified date range . Events from users who converted/installed your app before that date range are not included in your LTV data, even if the event occurred within that date range. For example, if today is June 1st and your chosen date is 3/1-3/31, your LTV report will show events and event values (to date) from installs that occurred in March , while your activity report will only show events that occurred in March, regardless of when the installs occurred. As you might imagine, the data differences can be huge, so you need to understand what they’re based on and how to analyze them.
The screenshots below show a comparison of this sample period, which leads to the conclusions:
LTV data is key to understanding mobile app marketing! LTV is a tool that shows the true quality of users from different sources over the long term and full lifecycle.
For simplicity, you can assume that all reports in Appsflyer are based on the LTV methodology , except for the Activity and Raw Data tabs.
The vast majority of sources, such as Google and Apple, display activity data that is not LTV data. If you want to compare data in any way, you need to start with AppsFlyer activity reports or raw data.
In the whole process of comparing data (by the way, the topic is not that important, but it is always one of the first to be raised) you need to pay attention to two issues:
If you understand the mechanics of collecting data, the principles of presenting it, and have a tracking plan in place, the reports themselves are really a piece of cake 🙂 The basic Appsflyer analytics dashboard, as I mentioned above, focuses on LTV data . This also shows Appsflyer's way of thinking about the app user as a long-term value.
After selecting the date, you will gain access to basic new user attribution data divided into organic and paid sources. You will be able to filter data by sources, campaigns or location. The system will automatically present the division of revenue per traffic source, as well as statistics on events from the most popular ones.
The table in the Appsflyer overview shows user acquisition data by source and the number of installs, re-attributions, and re-engagements. The more channels you integrate and tag, the better the quality of data you will get here . For example, connecting Google Ads and Appsflyer accounts shows data on impressions, clicks, and costs. You can add events that interest you to the table and analyze both the number of their calls and the number of unique users for these events. You can enrich the table with additional metrics, explore by specific campaigns or ad groups, and download the presented data to a file. Super simple and similar to the view you are used to in Universal Analytics.
It is worth adding that the standard Appsflyer dashboard is just an example , it does not always meet the appropriate conditions if, for example, you do not sell in the application. Hence, it is built on the basis of so-called widgets, which you can remove and add new ones, e.g. line charts, pie charts, tables, etc. Each newly created dashboard can be saved under a new name, downloaded or custom dimensions can be entered into it.
I described more about SKAN in a separate article, here I will only mention that Appsflyer, in addition to the dashboard, presents a separate tab of reports for iOS, for those users from iOS 14.5+. The tool presents here the LTV data of users assigned by SKAN in the configured SKAN measurement windows. In short, this place allows you to collect estimated iOS user data after configuring SKAN. If you want to analyze iOS data well, you should therefore look at both the data from the main dashboard and SKAN. However, is this data not duplicated? It is highly probable, the Appsflyer court introduced the so-called Single Source of Truth . This is a unified view for iOS, where the Appsflyer mechanism eliminates duplicate data on installations and corrects incorrectly assigned users.
This is possible by using data from multiple streams, including SKAdNetwork , ATT opt-in users, probabilistic modeling, and APIs for Apple Search Ads. This is meant to be a future view of aggregated iOS data with de-duplicated data in one place. Experience has varied, but if you send enough iOS campaign data, the reports will become increasingly accurate.
The cohort and retention panel groups app users by acquisition date . It's a good place to verify the user life cycle (LTV) by traffic sources, campaigns, and other media.
Cohorts can be built not only based on revenue, but also retention/engagement . The report allows you to group data in any way you want and colorize it using heatmaps. A great way to use one is to analyze the retention of an acquired user by medium/campaign over a given period of time. It often turns out that a higher retention rate beats a lower acquisition cost.
To date, there is still a report called "Retention", but it basically serves the same function as cohorts and will soon be deactivated, so it is not described in this article.
The activity report is one of the few places in Appsflyer where you can find data assigned to the event date, i.e. not related in any way to the LTV methodology. This place is often used either for comparison with other systems (does the data match?) or for current, short-term marketing analysis. So it answers the question of what sales we had in the app in March. On the other hand, it does not answer how the users acquired from January and February have converted so far. With these limitations in mind, it can be used better.
Available KPIs are:
Available event measurements in the app are:
Activity Report is a place for here and now analysis, very comparable to web campaigns. So if you don't expect LTV reports, but rather focus on current data, this is a good place.
The ability to create custom views is a distinctive feature of Appsflyer, something like the integrated Looker from GA4. Except that in Appsflyer it is faster and simpler, because it uses the idea of widgets.
I won't create detailed descriptions of each widget, but I will just point out that it is extremely useful to create tables with daily reports and selected metrics (very difficult to find them in other places), as well as add custom metrics such as ROAS.
The Events report is something like an inverted dashboard in Appsflyer, except that the starting point is events, not sources. The Events report shows LTV metrics in the app regarding user acquisition from different sources.
The Events panel is limited to events assigned to user acquisition campaigns , so it is not used for remarketing campaigns. I will also remind you that the date range in this panel refers to the installation date, not the actual execution of the event.
What can the event report be used for? I can recommend it for analyzing funnels and user transitions through the various stages of registration, purchase, etc. It is also a place where you can analyze unique users and the number of event calls per person.
The first step is to verify whether Appsflyer has built-in integration with the advertising system. Thanks to the integrations, you will be able to run attribution measurement, send feedback to partners, or gain access to cost data.
Most integrations consist of several steps:
The key to integration is configuring the so-called in-app event postbacks . This is information from Appsflyer to the advertising system about what events are to be sent with what names or values. It is important to maintain the correct naming, because some systems such as Meta Ads require standard events to optimize campaigns for them. The last stage of integration is to "capture" these events from the advertising system, e.g. Google Ads, and use them to optimize the campaign. To do this, you need to import previously mapped converters. This may take some time before they appear there.
Each integrated partner has its own naming convention in Appsflyer, for example Google Ads is googleadwords_int and you cannot change it.
As of 2021, Appsflyer has significant restrictions when it comes to tracking Google Ads iOS (ATT) traffic. Due to the lack of device identifier data:
To sum up, the data regarding iOS in Appsflyer with Google Ads will be significantly underestimated, so it is worth analyzing them in Google Ads itself.
If, in addition to paid campaigns, you deliver traffic to the app through other channels, such as a link on a website, social media profiles or via SMS, it is worth tagging these channels in advance. Otherwise, this traffic will go to the general Organic basket and further analysis will not be possible.
There are at least several ways to accomplish such integration:
Appsflyer is one of the most advanced MMP platforms for mobile app analytics . It focuses on LTV-based data presentation. It provides advanced reports related to attribution, SCAN usage and retention. Due to the huge number of integrations and tagging options, it is a great solution for analytics in one place.
Write to me about any matter related to app analytics at Appsflyer. I will help you conduct comprehensive training, obtain a special implementation offer or solve everyday problems.