Mobile application analysis of Android learning ability

1. What is mobile app analytics?

Mobile application is an important carrier of mobile Internet, mobile application Analysis refers to the use of mobile users, such as the basic data, data analysis by the service side, dig into the characteristics of user use, find product design deficiencies, discover the opportunities for operation and promotion, optimize product and operation strategy, Improve the quality of your mobile apps.

2. The significance of mobile application analysis

1. Monitor mobile app operation status
The most basic usage scenario for mobile app Analytics is to monitor the operational status of your application in real time. Through the data analysis , the Daily, weekly, monthly Report of the form of system monitoring. By analyzing and comparing the trends of these core data, using data to speak, we can help the operation and promotion staff to observe the operational status of the application in a multi-dimensional manner and identify problems in time.
2. Improve mobile app Promotion effect
Channel is the main way of mobile application promotion, different channels have the same user groups, according to the channel and the promotion location of data Analysis , choose low-cost, effective channels can effectively find the target users, improve operational performance, reduce the cost of promotion.
3. Discover mobile App product issues
Through the analysis of the application access path, custom event clicks, conversions and other data, to find out the reasons for user churn, the use of data to quickly locate product problems or strategic issues , and according to the standard mobile analysis ideas and methods to find a solution.
4. Optimize mobile app function and experience
As one of the main portals of mobile internet, the experience of mobile applications determines the user's fate to a certain extent, discovers product problems through 3, formulates optimization strategies , enhances the appeal and usability of the application.

3. Mobile App Analytics Metrics

Mobile applications are varied and the corresponding analysis indicators are different. In response to this situation, the mobile statistics Publishing white paper aims to generalize the ubiquitous indicators and definitions of various types of applications, and to make the operation of mobile applications more indicators.
Common indicators:
User Analysis Class indicator //imei+category information (uploaded at the start of the first activity)
1) New users: The first time in history to start the application users, need to follow the device number to weight.
Technical Note: The second installation does not count as a new user.
Implication: The more new users, the faster the application growth, the better the effect of promotion. In general, applications in the early stages of development when the proportion of new users is very high, as the market tends to grow steadily, the proportion of new users is declining gradually.

2) Boot count: The number of times the user opened the app during the specified time period. "One-time start" means that the user starts by opening the app to exit the app (or leave the app interface and into the background).
Meaning: The number of users is described from the scale of the application, and the number of start-up is measured from the visitor angle of access quality analysis indicators. If the user experience of an app is good enough and the user stickiness is high enough, the same user will start the app more than once a day, and the number of launches will be significantly greater than that of the visitors.

3) Active Users: Within the specified time range, the number of users who have launched the application, need to follow the device number to weight. Activity is the ratio of active users to the total number of users during a certain period of time.
Implication: Active users usually have a time range to do constraints, such as daily active users, weekly active users, monthly active users and so on. Active user indicator is an embodiment of the application of the user scale, but also a measure of the application of the quality of the most basic indicators, combined with retention rate, wastage rate, the use of time and other indicators can also reflect user stickiness. The indicator can also measure channel quality and troubleshoot channel cheating.

4) Retained users: The user who is still using the program after a period of time (T2) in the new user within the specified time period (T1). where T1 and T2 can be set according to the actual situation of the application itself.
Implication: Retained users are mainly used to measure the application to the user's attractiveness, user stickiness of the application, channel user quality and delivery effect. The usual retention indicators were retained on the following day, retained on 3rd, and retained on 7th.

Use the Behavior class indicator //imei+category+application time+activity time to upload when all activity is destroyed
1) Length of use: The amount of time the user stays on the application. It is divided into the average duration of use and the duration of a single use, and the average length of use is the average of all access times for a given period of time for all users.
Implication: By considering the user's dwell time on the application, we can see whether the application content attracts the user, whether the application quality is qualified, and whether the user who arrives at a promotion channel is a deep user, in order to judge the channel quality.

2) Usage Frequency: The number of times the same user initiated the application during a certain period of time. If in a day, the same user to effectively start 5 times, then the user's daily use frequency is 5 times.
Implication: the use of frequency and the number of days start similar, just from another angle to measure user stickiness, an application usually the higher the user stickiness, then the average user frequency is higher.

3) Use interval: Use interval is the time interval that the same user launches the app two times, for example, when a user starts the app for a period of 2 days between the first launch and the second launch, the user's use interval is 2 days.
Implication: The use interval also reflects the user stickiness of the application from the side, usually the shorter the use interval, the more users rely on the application, that is, the higher the user stickiness of the application. The timing and frequency of the push message can also be determined accordingly.

4) Depth of access: The cumulative number of pages a user has reached during a single start-up application is considered a user's depth of access, such as when a user has visited 12 pages from the start of the app to the Exit app, and the user's access depth is 12.
Implication: Theoretically speaking, the higher the access depth, the better the application quality, the stronger the user's dependence on the application.

Channel Operation Index //imei+category+location
1) User Acquisition Cost: The cost of acquiring a new user, which is the marginal cost that the user gets, the cost of acquiring a new user is reduced as the number of new users becomes higher, and vice versa.
Implication: User acquisition Cost is the most important ROI (return on investment) indicator in the application promotion process, if the user acquisition cost of a channel is obviously higher than other channels through statistical analysis, then it can abandon the channel, and put the main promotion funds and resources in the channel that the user gets lower cost. In order to get the most new users within the unit resources.

2) Target event conversion rate: After setting the conversion conditions and conversion results of an event, when the conversion conditions appear, there will be some conversion results, the target event conversion rate is the number of conversion results/pre-conversion events. For example, to set all start-up users click to complete the registration as a full conversion event, click Register is the conversion conditions, become a registered user is the result of conversion, if every 100 access users 50 people to complete registration as a registered user, then you can say that the target event conversion rate of 50%.

Implication: Tracking statistics for specific events is a supplementary function in mobile application analysis, but it can realize the data that the general function cannot count, such as the proportion of registered users, the proportion of paid users, etc. The target event is often accompanied by a custom event feature that fully meets the personalization and customization needs of mobile developers.

application Quality class indicator //imei+category+ system + System version + mobile phone model +memoryinfo+poweinfo+trafficinfo+errorinfo
1) Number of errors: in the specified time period, the total number of occurrences of abnormal exits is applied.
Implication: In general, application errors focus on two kinds: The crash or the strong retreat, such a phenomenon will seriously affect the user experience, so the higher the number of errors occur, the user experience is worse. Reducing the number of errors is an important indicator that developers should always be concerned about.

2) Error Rate: The rate at which an application has an error (number of errors/number of starts) during the specified time period.
Implication: Product quality is a cornerstone of the development of the application, and the site, mobile applications once distributed can no longer recover, the application of high error rate will reduce the user's product trust and reputation. By monitoring the quality of the application, locating the error code and releasing the repaired version in a timely manner, this problem can be effectively remedied.

4. Mobile App Analytics Process

The mobile analytics process is a process of discovering, analyzing, and resolving problems. Leveraging the prescriptive analysis process in the mobile app's operations avoids confusion and makes it easier to optimize applications more clearly and efficiently.
clear purpose, establish planning
1) Profitability: Directly increase the profitability of the application, such as the application of the game class
2) Brand: Continuously strengthen mobile application as the influence of mobile advertising media
3) Cooperation: to serve users or other applications to import traffic, such as 91 mobile phone assistants, etc.
Select systems, deploy Apps
Features: Simple development, comprehensive data, complete function, reliable system
such as: Baidu Mtj.baidu.com, friends Union, etc.
measuring indicators, analyzing phenomena
Here's the indicator that the analysis begins to talk about
tracking issues, continuous improvement
This depends on the vitality of the individual.

5. Mobile Application Analysis Ideas

The thinking of mobile application analysis mainly includes microcosmic subdivision and macroscopic trend.
Micro-Segmentation is the focus of analysis on the details of the insight and grasp, such as the number of new users, daily activity, page bounce rate, user stickiness, the conversion of custom events, in detail to consider the mobile application of specific indicators and KPI performance, so as to achieve specific problems specific analysis.
Among the statistical analysis objects of mobile application are: Channel source, audience attribute, user behavior and terminal equipment.

macro Trend is to control the mobile application from the overall data on the year-over, the chain and other trends, the core indicators in yesterday, last week, the change of last month and so on. This is from the macro analysis of the trend of application development, today through the application of internal and external data for comparative analysis, clear their overall development status and position in the industry, so as to help developers to specify macro development strategy.
The statistical analysis objects of mobile application include: Time series, yoy, chain, fixed base ratio.

So it can be said that macro trends and micro-segmentation combined with the use of "Think big, small start" operation Management, to achieve "targeted, well-aware."

6. Mobile Application Analysis method

Audience Analysis-Knowing the user can make a more popular application
The audience analysis includes several subdivision directions that are described in detail:
1) New Active user analysis
The new high ratio indicates that the future development potential of the application is greater, and the higher the daily activity, the greater the user stickiness of the application.

2) Analysis of retained users
The most common is the first use of retention analysis, which refers to the use of the app's users, the first time after the use of the same period of time after the use of the same interval. It is divided into the concepts of daily retention, weekly retention, and monthly retention.

3) Population attribute analysis
The user is a real person, then the user community must have its specific demographic attributes, including gender, age, education, interest, industry and so on. Therefore, for an application, its population attributes must have tendentious characteristics, such as the application of a user group with a high degree, more men, computer industry, like reading and other characteristics.

4) geographical distribution of users
Not explained.

using behavioral analysis-well known, optimizing the application process and the basis of page design
In mobile application analysis, audience analysis is the static analysis of users, the main grasp of user static properties, and the use of behavioral analysis is the dynamic analysis of users, focusing on the user's use of dynamic Data. On the basis of understanding the users ' attributes, this paper analyzes the usage behavior of these users in the application, and can complete the deep understanding of the users more comprehensively and systematically.

For developers, mastering the behavior and participation of users can better guide them to optimize the application process and page design.
Usually user behavior analysis is also called participation and use analysis, including the following aspects:
1) page Access Path analysis
2) Access Depth analysis
3) Use interval analysis
The usage interval is the interval of time that the same user launches the app two times, within the selected time range.
4) User stickiness analysis
User stickiness is a comprehensive dimension of the analysis, in the previous article has also been mentioned, many indicators can be reflected from the side of the application of user stickiness, the most direct measure of user stickiness is the use of time and frequency of use. Typically, the longer the user is used, and the higher the frequency of use, the greater the stickiness of the user.

channel Source Analysis--strategist, channel monitoring is the tool to measure the effect of promotion
There are many promotion channels for mobile apps, including App Store, app market, in app promotion, etc., developers have invested a lot of money and energy in channel promotion, but the effect is not good.
The following is a detailed explanation of the Channel Source Subdivision feature for two operating system platforms.
1) Android Channel URL Download source breakdown
Android apps for a variety of reasons, distribution in the country relies largely on unofficial application markets and websites, while Google Play's domestic share is almost negligible.
The operation is to the channel promotion of the original connection URL encapsulation, the developer can easily add new channels through the background, and then can track the different channels and different promotional locations brought by the download volume report.

2) iOS channel URL download source breakdown
Added the App Store upstream source segment.

End-Version analysis-know-how to improve application capabilities for different devices
A developer's gospel that counts device adaptation issues for mobile applications, such as what models are not known about the mainstream of the target user base, or how their operating systems, networking, resolution, and other data are. and the terminal version analysis function can let the developer know the device terminal and the distribution of the version in the user group, thus greatly reduce the development load, improve the development and adaptation efficiency.
1) Terminal analysis
including equipment model, resolution, operating system, networking methods and other indicators, each indicator is equipped with a histogram chart, there are detailed data reports.

2) version distribution
A reasonable version distribution and management strategy, can greatly improve the user experience, because too many versions and too fragmented distribution, increase the developer's management costs, but also not conducive to the new version of the promotion.

custom event and error analysis--a powerful, personalized and intelligent application statistical analysis
1) Custom Event Analysis
Customize the information you want to see

2) Error Reporting analysis
Because the mobile app download is installed directly on the user's device, so many errors can not be found in time, this is the cause of the application of the bug is difficult to open the development of the cause of discovery.

Citation: Http://mtj.baidu.com/web/dashboard
Baidu Mobile Statistics

Mobile app analytics for Android learning capabilities