Geekly Articles each Day
📜 ⬆️ ⬇️

Mobile App Metrics Overview



So, you have published your first application in the store. The first downloads started, and now is the time to start removing metrics in order to analyze them and identify possible weak points. Analytics - the most important tool in the world of mobile applications. It allows you to understand the psychology of the user, to understand how he interacts with the mobile application, and as a result will help make your child better and more profitable.

There can be a lot of metrics, and usually their set depends on the specific application. But there are a number of key indicators that need to be monitored, regardless of the nature and scale of your project. These include:


Let's look at each item in more detail.
')

Application Installation Source


A very important metric that allows you to understand the effectiveness of a particular advertising channel. You can simply track the advertising channel, the principle is the same as in the case of the transition to the website: in the link leading to the app store, special labels are inserted that are unique to each of the advertising channels. After installation, the application reads these tags and records the source. Further, this source is displayed in the analytics system that you use.

User retention


It uses a variety of metrics. After the user has installed and launched the application, he assesses whether he likes it. If not, he will immediately remove it or close it and forget it. But if the application was like, then after a while the person will launch it again.

To assess user appeal, metrics are most often removed:


All three metrics are removed daily, each time the application starts, the current date and the date of installation are compared. Analysis of the dynamics of changes in each of the metrics will also allow you to understand the reaction of users to certain changes you make to the application. For example, a 1-day retention level usually indicates how users react to your application interface. And if this indicator began to decline, then first of all it is necessary to check what is wrong with the interface.



The next important daily metric is the increase in the number of new users . Moreover, it is recommended to monitor the change of this parameter when conducting advertising campaigns, placing review articles, entering into partnership agreements, etc. In this case, the metric acts as the effectiveness of all these movements. It is advisable to impose on the graph of the number of new users not only dates, but also installation time, which will help to more accurately assess the role of the promotion and advertising measures you take. It is also often useful to evaluate the dynamics depending on the geographical separation of users, as well as separately for different user segments.

If the dynamics of growth will be negative, then you need to actively engage in promotion and advertising. We will tell more about this in a future publication.



The number of unique users during a certain period


So, you managed to achieve a more or less sustainable growth of the audience, the project is warmly received by users and is gaining popularity. It's time to think about the degree of user activity: how many people launch your application per day? And a week? Per month? And we are talking about unique users. Three metrics answer these questions:


In essence, each of these metrics is computed from one common database, in which statistics is accumulated for all application launches. The uniqueness of users can be determined, for example, by assigned ID or login / password pairs.



You can also calculate the derived Sticky Factor metric = DAU / WAU or DAU / MAU. Its name can be translated as “degree of stickiness”. It describes the regular use of your application during the week or month, that is, it allows you to assess how much people like your application based on the frequency of use. If all users run the program every day, then DAU will be equal to WAU and MAU, and their ratio will be 100%. But this does not happen, and therefore Sticky Factor allows you to assess how often people access your application during a week or a month. It is logical that the decline in these indicators - an unpleasant signal, talking about the cooling of the audience.

Session


A session is the time that a user spent in a mobile application from the moment of launch until the end of its use. In relation to sessions, two metrics are usually removed:


However, you should not chase after the high values ​​of this metric, because it strongly depends on the type of your application. For example, for games, this indicator is quite critical, and the more it is, the better. And for applications, widgets or fitness trackers, this figure will be insignificant, since for the most part they work in the background. It is much more important to know which screens the user visited during the session . Thanks to this metric, you can determine the sections of your application that are most interesting to users. And at the same time, you will find out which ones can be completely removed and not to be engaged in their development in the future.

A very useful metric is on which screen the user session ends . This indicator is important, for example, if you have authorization in the application. It often discourages users, especially if the application does not allow viewing content, but first requires a login and password. In this case, the session will most often terminate on the registration screen. If you add some content before logging in, then thanks to this metric you will immediately see the result.

Another example: if you have a product order form consisting of 3-4 screens, then this metric will show at which step most users leave the application. As a solution, reduce the number of steps, optimize their order or design.

Interactions with interface elements


Trying to raise the values ​​of certain metrics, very often you have to adjust the user interface and change the functionality of the program. You can evaluate the effectiveness of these steps using A / B testing (real-time testing, when a group of users is offered one version of the functionality / content, and the rest of the users are offered a different version). In our case, testing implies rolling out a new version of the application with a modified UI for some control group of users. The rest continue to use the current version. And we register how the control group reacts to innovations by removing interaction metrics with the application interface: for example, which of the two buttons gives a higher purchase conversion, where better to show the popup asking for feedback about the application, etc. You can also use third-party services for A / B testing, for example, Apptimize , Optimizely , Mixpanel .

With the help of collected statistics, you can also find out how much this or that application's functions are in demand, how many users interact with the application without connecting to the network, and much more.

Finance


This is one of the most interesting and important groups of metrics. If you plan to make money with your application, then you need to pay close attention to registering these metrics and controlling the dynamics of their change.

The first thing that comes to mind is the total amount of payments for the period, Gross . However, keep in mind that this is a gross income, from which you still have to subtract the share of the store through which you distribute the application. But after deduction, we get the Revenue metric, which reflects the amount credited to your account.



Suppose your application itself is free, but some of the content is available only for money - you distribute it in in-app purchases. To develop the application and increase revenue, we need to know how many unique users are paid during a given period . For example, how many people bought game tokens, golden shells, more powerful spells, access to advanced analytics, beautiful design or other paid delicacies offered by you per month.

The following metric is derived from the previous one: what proportion are the payers of the total number of unique users (for the period) Paying Share . Our unattainable ideal is 100%. Although in reality everything is usually much more modest. If this indicator starts to fall, then users are already fed up with the existing paid content, and it is time to either diversify it or play with discounts. On the last point there are many different tactics. For example, you can give discounts on weekends and on holidays. You can create a stir, temporarily bringing prices down, and as soon as the number of downloads increases significantly, return the prices back to their previous level. You can give discounts on coupons, you can offer to perform some simple quest. Another option: "the first discount of 5,000 people who downloaded Ivan Kupala on the night of". If there are other applications with paid content in your portfolio, you can use package discounts when downloading two or more of your products. In general, there are quite a few options for using discounts.

In addition to the number of taxpayers, we are also interested in the specific number of payments per user, Transactions by User . This metric is calculated by the formula:

TBU = T / PU, where T is the total number of payments (transactions) for a certain period, PU (paying users) is the total number of payers for the same period.

If TBU> 1, then some users made more than one purchase.

The following important ARPU and ARPPU metrics are:


Speaking about the profit received from users, we should not forget about how much their attraction costs us. In the end, the first must be greater than the second, otherwise what is the point in all this? As a metric, the cost of one application installation (CPI, Cost per Install) is used here . Calculated by the formula:

CPI = A / I, where A is the cost of advertising, promotion and marketing, I is the number of application installations.

This metric can be calculated both for the entire lifetime of the project, calculating the current cost of attracting the user, and for certain periods, determining the effectiveness of specific advertising campaigns or measures to promote the application.

And we conclude our review with the LTV (Lifetime Value) metric - this is the specific user profitability throughout the entire period of using the application. There are many ways to calculate LTV, but first you can use the following formula:

LTV = ARPU * Lifetime, where Lifetime is the average duration of using the application from the first launch to the last. For example, if a user first entered the application on January 1, and the last time - on August 15 and did not use them anymore, then for him Lifetime is 7.5 months. By summing Lifetime for all users and dividing them by their total number, we get the average value of this metric, which will be used to calculate LTV.

Note that when calculating LTV, the Lifetime multiplier should be a multiple of the period for which the ARPU is calculated. If you took ARPU for a month, then Lifetime will be measured in months, not days or weeks. Let's say if your application has a monthly ARPU of $ 5, and Lifetime is 3 months, then LTV = $ 5 * 3 = $ 15.

This metric is one of the key parameters for evaluating the effectiveness of your project. If LTV is less than CPI, then the project is unprofitable without any “if” and “let's take a different look”: you spent more on attracting the user than you received from him for all the time that he used your application. Therefore, LTV must be constantly monitored and immediately respond to the downward trend of this metric. Obviously, you can increase LTV using one or both multipliers, achieving an increase in the average profit per user per period and / or an increase in the average duration of use of the application. For example, you can reduce the outflow of users, increasing the attractiveness of the application; reduce borrowing costs by choosing more efficient channels; increase the cost of purchases by raising prices and stimulating the need for paid content.

Finally, we want to give an example of metrics for two popular games: Mobile Strike and Clash of Clans. The summary data on versions for Android and iOS in the USA are given. If you make mobile games, you can focus on their metrics, as on the top products in this class of applications:


More in the screenshots
Number of downloads:



Weekly number of unique active users (WAU):



Ratio of daily and weekly active users (DAU / WAU):



Daily Profit:



The matrix of two indicators - 30-day retention and frequency of use per week - for different categories of applications according to the Flurry analytics system:





About analytics systems


There are quite a lot of them, but Google Analytics , Flurry and App Annie are the most popular among mobile application developers. For the first time you will be more than enough of their capabilities. All tools are offered to developers of SDK for iOS, Android and Windows Phone, which are easily integrated into the finished project. Consider more.

Google Analytics


Google Analytics is a very powerful and completely free tool for removing metrics and subsequent analysis. Initially, it was created for web applications and websites of various levels of complexity, therefore, it is not very convenient to use mobile applications, but it copes with basic tasks.
Interface


Mobile developers are especially interested in the "Real-time" section. Here you can see online the number of users of the mobile application, as well as events that you will be tracking.

In general, Google Analytics is most suitable for programmers and indie developers.

Flurry


This tool was originally designed for mobile applications, so with them it is more convenient to use. Like Google Analytics, Flurry is free to use. The interface does not look too cluttered, this is a distinct plus compared to GA.
Interface


Flurry focuses on tracking user behavior, so most out of the box reports are somehow related to this direction.

This tool is more suitable for marketers and analysts.

App annie


This service has free basic functionality, which is enough for novice developers. But if you want to shoot a wider set of metrics, then you have to pay . Classic interface: on the left is the navigation bar, and the content is conveniently arranged.
Interface


In general, this service can be equally useful for developers, and for marketers with analysts.

Google Analytics and Flurry provide all the necessary basic tools for monitoring mobile applications. Free App Annie functionality is somewhat limited, but they have two paid programs with much more features - for medium-sized companies and Enterprise.

Google AnalyticsFlurryApp annie
Analysis of download sources+paid
User analysis++paid
Analysis of various platforms+paid
Conversion map++paid
Analysis of the effectiveness of advertising and attraction++paid
User behavior analysis++paid
Financial performance++paid
Active users++paid
Cohort analysis++paid
The ability to create a panel with its own set of reports++
Top developers+
Top applications by category and site+
Revenue top applicationspaid
Retention of top applicationspaid
Using top applicationspaid
Audience top applicationspaid
Marketing top applicationspaid

Summary


Mobile application analytics is a very important part of the project life cycle. For individual developers and small studios, it is vital to keep abreast of their projects, nurture them and immediately react to negative signals, manifested in the deterioration of metrics when the project starts and global milestones every hour.

The described analytics systems are only part of the arsenal of tools that facilitate the work of many studios and independent developers. Today, the creation of successful applications requires the acceleration of the development process, the use of convenient and functional tools. Based on this, we are developing Scorocode, turning it into a useful, and for someone, an indispensable tool for developing mobile applications.

Good luck to you development and high revenue.

Source: https://habr.com/ru/post/308438/

More articles:


All Articles