The Ultimate Guide to Mobile Analytics and MMP

Once a mobile app has been created, you should carry out a standard set of actions to ensure that its intended users can find it organically through the search bar, then download and install it on their device.

You can constantly expand the semantic core, improve the metadata, change the screenshots, run various tests, try out creative ads, introduce new monetization systems, etc., but what good is any of it if you don't understand what results all this produces? So the logical question arises: how do you evaluate an app's key performance indicators (KPIs)?

It is essential for both developers and marketers to understand the strengths and weaknesses of the project. The detailed analysis allows you to develop a strategy for further growth and improvement of the product. And this is where mobile analytics comes in. This article from RadASO is about exactly this issue.

What are mobile app analytics and MMPs

Mobile analytics is the process of analyzing an app's key performance indicators to understand all events within it.

Mobile analytics tools, also known as MMPs, help assess the necessary KPIs.

A mobile measurement partner (MMP) is a tool or service that collects and organizes mobile app data to measure app performance across all channels. These can be advertising marketing channels, media sources, and advertising networks. It collects data for engagement and conversion analysis to better understand the users and the product, thereby increasing the app's profitability.

The main objectives of mobile analytics are:

1. Pinpointing marketing channels and identifying traffic sources within analytics to identify what conversions are occurring and which are the most effective.
2. Tracking installs across different channels.
3. Measuring the app's profitability if it has in-app purchases.
4. Identifying events in the app.
5. Deep linking – attracting users to specific product pages.
6. Analyzing user audience behavior.
7. Monitoring the app's stability.

The main types of mobile analytics

There are three main categories:

1. External (marketing) – identifies the app installation sources and shows the effectiveness of each.
2. Internal (in-app) – tracks all user actions, i.e., analyzes their behavior.
3. Performance analytics.

External mobile analytics

The objective of external analytics is to track the performance metrics of different sources to evaluate the effectiveness of your advertising campaigns. Yes, most users find the app through search, but it's challenging to stand out among the thousands of other apps purely organically.

An app's success also depends on marketing campaigns, whose main purpose is to attract those users who will install the app on their device, actively use it, and generate revenue.

Working with multiple advertising sources is one of the most popular ways to promote an app. Therefore, it's important to understand which campaigns are the most effective. To do this, external mobile analytics should be put in place to track key metrics, such as:

• installs (not only their number but also the path leading the user to take that action);
• views;
• app launches;
• clicks;
• purchases;
• registrations;
• shares – on which platform your app is linked;
• invites – the number of invitations, who invites whom.

Using this data, it's possible to determine which ads a user clicked on before installing the app and optimize advertising campaigns based on these insights.

Also, marketing analytics can identify precisely which publisher is influencing conversions. The problem is that it's difficult to distinguish which publisher the install came from when advertisers work with several publishers.

Often, the same user clicks on two ads, and then two ad networks must be credited for the app install – resulting in double payment for one app install. To avoid this, marketing analytics matches a new user to a specific click on an individual ad, so you can be sure which publisher the app was installed from and thus be able to optimize the cost of your ad campaigns.

Internal analytics

In-app analytics scrutinize all user actions within an application, including

1. Device profile (its type, manufacturer, and operating system).
2. Demographics (user location, gender, approximate age, language).
3. Determining whether the user is new or has used the app before.
4. User behavior in the application (this includes all events in the app – taps, purchases made, sections viewed, functions used, levels completed in games, i.e., all buttons pressed).
5. Number of installs and downloads.
6. Duration of sessions.
7. Conversion rates and profit margins.

You can use the collected data to optimize user interaction. For example, you might find that most people use your app on a tablet and optimize it for that type of device.

If it's a game, you might notice that a large number of users leave the game at the same stage. You can assume that the difficulty level there is too high. So, making it easier to progress to the next level will help retain players longer.

In addition, with some mobile analytics tools, you can optimize and personalize the subscription page (change prices, trial periods, and promotional offers), thereby increasing the number of subscriptions. Subscription screen testing is done without having to release a new version of the app, as these A/B tests can be run directly from the app itself.

As an example, The Meditation App managed to increase its profits by 30% with the help of Adapty.

Performance Analytics

The main performance indicators of the application are:

• uptime;
• response speed.

If an app's performance is unsatisfactory, the user is unlikely to use it. Therefore, performance analytics enables you to identify the causes of technical issues and prioritize solutions timeously.

This type of analytics collects the following data:

➢ API delay;
➢ operator or network delay;
➢ data transactions;
➢ failures;
➢ uptime;
➢ technical errors.

How does MMP work?

MMPs offer an SDK* (software code part) for integration into an app.

A mobile measurement partner uses multi-sensory attribution to identify each interaction point that influenced a user's decision to install an application that came from a specific source. This tool tracks clicks, app installs, and all further actions directly in the app.

Examples of MMPs:

• Appsflyer;
• Adjust;
• Firebase, etc.

When a tracking system is integrated into the app, the process follows this algorithm:

1. A user clicks on an advertisement and is taken to a referral link in the analytics system.
2. The referral link redirects the user to the store, and information about the link's source is stored, i.e., the interaction with the ad.
3. The user downloads and uses the app, and the tracking system identifies information about the user, their device, the source of the click, etc.

Following these steps, when the app is first opened, the MMP tool receives the following metrics:

• Advertising ID – the identifying device (each individual smartphone or tablet).
• IP address – the address devices use to communicate with each other over the internet.
• User agent – detects the user's browser and operating system.
• Timestamp – the time when the ad was clicked.
• First install – records the timestamp of when the app was opened for the first time.

This is how the MMP detects whether the user is new or has used the app before.

The app installs are then matched against ad interactions. Conversion is assigned to the publisher from which the app install occurred so as to differentiate mere impressions from users who decided to take the next step and initiated an action that you've set to count as a conversion (such as installing the app) and put them in a separate group than those who are just browsing.

Each MMP system has a different approach to detecting the source of traffic. For example, let's look at the attribution scheme in Adjust:

In AppsFlyer, attribution is based on the following scheme:

In mobile analytics tools, app installs are recorded and tracked from the moment the user first opens the app.

Advertising sources record the interaction time, while stores record the download time.

This allows the professional to properly analyze data and allocate marketing spending, giving preference to the sources that generate more revenue. After all, mobile analytics not only reveals the source of traffic but also tracks all further user actions. What also matters is whether a particular user has made any profit as a result.

Why an ASO specialist needs mobile analytics

Of course, the indispensable tools for the ASO specialist are the App Store Connect and Google Play Console mobile analytics systems. These services allow us to view app statistics on impressions, installs, conversions, profits, etc. Each store has its own peculiarities, so let's look at them separately.

App Store Connect

App Store Connect is an easily integrative tool available for all iOS apps.

The Overview tab shows app performance data in the form of percentages, charts, and graphs.

Main indicators:

• Impressions – number of times the app has been viewed in the App Store.
• Product Page Views – number of product page views of the app on iOS 8 devices and above.
• Conversion Rate – number of app installs in relation to the number of impressions on unique devices.
• Total Downloads – number of first and second downloads.
• Proceeds – developer's profit on sales minus Apple's commission.
• Proceeds Per Paying User – total proceeds divided by the number of paying users.
• Sessions Per Active Device – refers to how many times an app was used on a device for at least two seconds.
• Crashes – number of crashes.

In the same tab, you can sort data by territory, traffic source, or device.

App Store Connect distributes traffic to the following sources:

• App Store Search – organic traffic, but also includes traffic from Search Ads.
• App Store Browse – traffic by feeds, categories, and selections.
• Web Referrer – traffic from third-party sites.
• App Referrer – traffic from apps (e.g., Facebook, Instagram ads).
• Unavailable – the source type is unavailable because customers may download the app before App Analytics starts tracking the source attribution.

The Metrics tab allows you to track all metrics using filters for convenience, for example:

• by download date;
• app version;
• device type;
• page type;
• embedded events;
• region;
• source type, etc.

The Retention tab allows you to view app usage – the percentage of users who continue to use the app for a given period.

However, App Store Connect is by no means perfect; an ASO specialist will find it has its fair share of drawbacks, for example:

1. Incorrect traffic attribution – displays and app installs from ASA (Apple Search Ads) fall into organic (Search).
2. The App Store has no separation of purchases (Proceeds, Sales, In-App Purchases, Paying Users)* coming from the Search Ads source; all this data goes to Unavailable along with the others (whose sources could not be established).

When estimating sales or proceeds, around 30-50% of the data will be in "Unavailable".

Profits are shown as follows (Search Ads are not shown separately):

Google Play Console

This is what the Google Play Console interface looks like:

Using filters, you can view shop page visitors, acquisitions, and conversion rates by country:

• App Install State – whether the user currently has the app installed on their device or not.
• Country/Region – the country or region where the user's Google account is registered.
• Language – depends on the language setting on the user's device.
• Search Term – the query that the user searched for before going to your app listing page.
• Store Listing – the shop record that the user visited.
• Traffic Source – how the user got to your app listing page.

Google Play Console also allows you to track failures and monitor the app's rating:

As for the disadvantages, Google Play Console has its share, too.

Google UAC* can hit different sources:

• ads in search go into searches (Google Play Search);
• ads in selections end up in selections (Google Play Explore);
• YouTube ads are added to referrals (Third-party Referrals).

To identify the referral resource and specific ads in more detail, an additional mobile analytics tool needs to be integrated into the app.

Traffic in Google Play Console is displayed by the following sources:

• All Traffic Source – total number of users who installed the app from its Google Play app listing page.
• Third-party Referrals – number of users who installed the app by going to its page from any source other than Google Play.
• Google Play Search – users who installed the app via Google Play Search.
• Google Play Explore – users who installed the app via Google Play but not via search (traffic by feeds, categories, and selections).

Purchases in the console by source are not separated at all; there is no option to apply such a filter:

This is also a problem because it's not only important to know where users came to the app listing from, but more importantly, what actions they then took, whether they made any purchases or subscribed, or brought in revenue for the developer.

How extraneous mobile analytics services can help

With the help of mobile analytics services, you can resolve the problems described above.

For example, this is how the traffic distribution in AppsFlyer looks graphically:

This way, all indicators from different sources can be tracked in tabular form:

It is also possible to track which ads from the same source are bringing in which traffic:

There is also a split by source for profits section:

As for Adjust, this tool presents the data in the following way:

How to choose a mobile analytics tool?

Before selecting a complementary mobile analytics service, you should understand what key metrics show the app's effectiveness and what exactly needs to be tracked. There are a lot of parameters you can analyze, for example, you can aggregate data on costs, segment the audience, track user behavior, etc. This is why it's important to understand the main objectives when choosing a tracking system.

If we're talking about the needs of ASOs and separating organic performance from advertising and, therefore, accurately determining profit from different sources, then any analytics system will do.

However, when it comes to tracking on a larger scale, it's better to integrate a tool that meets the needs of marketers and application developers alike. Therefore, let's highlight the main parameters we will rely on when choosing a service:

• Self-Reporting Networks (SRN) – self-reporting networks report events (e.g., app installs or internal events in the application) through an API.
• Cohort Reports – cohort reporting allows you to see at what point and from where the user removed the application.
• Impression Tracking – if the user didn't do anything other than installing the app after viewing an ad, the app install will be attributed to that advertising source.
• Audience Segmentation – allows encrypted user data to be sent to advertising networks. Owing to this, it is possible to create personalized advertising campaigns.
• Custom Dashboards – ability to customize the analytics interface itself.
• Custom Reports – ability to selectively specify parameters for reports.
• Advertisement Cost – display advertising costs for each channel.
• DAU/MAU – number of active users, which includes DAU (Daily Active Users) and MAU (Monthly Active Users). By dividing these figures, you can determine the level of user loyalty as defined by the Stickiness app.
• Raw Data Export – allows you to obtain maximum details on all events and cohorts.
• API Reporting – ability to fetch data from the tracker to your server programmatically, without having to do it manually. 

After comparing the services based on these parameters, we get the following results:

AppsFlyer

AppsFlyer is one of the most popular mobile analytics tools, so let's take a closer look at it.

Customizing

1. Account setup. It is advisable to set up an account right away for the same person who will be the main user of the tool, as all messages will be sent to the email address specified during registration. Once registered, you can add additional users and configure their roles and access certain features.

2. Adding an application.

Next, you need to select the application platform and its current status and specify the URL:

Even if the app is not yet released in the store, you can still add it. As soon as the app is available, it will automatically display the "Available" status in AppsFlyer, and you can measure the effectiveness of your campaigns.

3. AppsFlyer SDK integration. At this stage, it is necessary to involve the app developers.

Before installing the SDK, it's important to understand which internal features you want to measure. By doing so, you'll not only be able to assess the effectiveness of your campaigns, but also understand the quality of users coming from different sources, i.e., which ones bring in the most revenue.

AppsFlyer has an in-app event generator where you can get a list of recommended events by selecting an app category:

After selecting the event, the resulting code should be sent to your mobile developer, who will integrate it into the app:

You also need to understand:

• whether you will be using deep linking (deep links are those that lead to specific sections of your application);
• how you would like to receive data.

It is essential to ask yourself these questions from the very beginning, as the appropriate configuration needs to be set up in the SDK.

4. Integration of the advertising sources you run your campaigns through (e.g., Google Ads, Facebook Ads, Apple Search Ads or your own media sources).
5. Testing of integration and settings.
6. App release and data collection.

AppsFlyer works with all traffic sources and has a handy dashboard. There is a free plan for 12,000 conversions. Subsequent rates are negotiated individually.

Key features of Appsflyer

1. Traffic measurement – the tool's dashboard separates organic and non-organic installs, and their internal events.

2. Comparison of both user attraction and retargeting data:

3. Filtering of indicators, e.g., by media resources, countries, publishers, attribution types, and sources.

4. Comparison of key KPIs: impressions, clicks, installs, conversion rate, cost, profit from different sources:

5. Return on investment (ROI), effective price per install (Avg eCPI), average revenue per user (ARPU), session duration:

6. Real-time tracking of metrics:

7. Focusing on the sources that bring in the highest number of users:

8. Comparison of daily ROI by grouping users by geo or media source:

9. Ability to customize dashboards and reports, selecting only the metrics that matter to you.

10. A convenient way to download all the reports you need and schedule them via email:

Adjust

• It has the same basic features as AppsFlyer but has the advantage of letting you customize reports (AppsFlyer offers this service for an additional fee).
• There is no free plan, but there is a demo mode where you can test the basic functions.
• Unlimited lookback window, i.e., data is saved for the entire period of use (AppsFlyer saves data for 90 days).
• Cohorts on all sources have no time limit.
• There is an anti-fraud mechanism that can track different types of forged traffic from various sources.
• Easy settings for reattribution.
• Ability to configure the attribution window at the program, partner, campaign, and ad group levels.

Disadvantages include:

1. Relatively high cost.
2. Initial plans do not come with a personal manager service.

Conclusion

Mobile analytics is an indispensable tool for both app developers and ASOs. Data analysis can identify project strengths and weaknesses and help you draw conclusions on further actions to improve efficiency.

Why MMP is useful:

1. You can see traffic (impressions, clicks, installs, purchases and profits) from a specific source (Apple Search Ads, Facebook Ads, YouTube, website ads, etc.). Since all the data is displayed in one place, you can save time on analysis.
2. It is possible to track not only the source where the user came from but also the specific ad they clicked on before installing the app.
3. Knowing how much revenue each source and each advertising post, banner, etc., bring in, you can correctly allocate the spending budget because you can see which advertising campaigns are effective and which are not, thus allowing you to prioritize them properly.
4. You can study the app's target audience – which users use it: their gender, age, region, which devices they use most often, etc.
5. Track user behavior within the app – which features are popular, after which section they decide to leave, and so on. By doing this analysis, you can take steps to retain users as long as possible.
6. View conversion rates, user retention, and churn rates.
7. Check the app's stability to reduce bounce rates.

This data will help you develop the most effective mobile marketing strategy to make your app profitable.

Text localization: Kateryna Kalnova, RadASO.