Different users have different preferences and different product requirements. To better cultivate user stickiness, it is necessary to implement individual policies and refine operations. Audience segmentation is one of the important methods of refined operations. Audience segmentation is the process of grouping users with the same or similar characteristics into a group based on user attributes and behavior data, and formulating differentiated operating strategies for users in different groups to improve user retention and conversion.
Audience segmentation application scenarios
In our daily work, we often encounter the following scenarios and problems:
1. The overall user retention rate has shown a downward trend. Which users lowered the average retention?
2. Every day there are users who receive coupons or points, but do not use them. How to find out these users and prompt them to use them as soon as possible?
3. How to conduct hierarchical operations for users of different regions, models, ages, and consumption levels?
4. How to trigger contextual messages based on user behavior and interests?
5. Is it possible to trigger an upgrade reminder in multiple scenarios for users of the old version without releasing the version?
……
All of the above can be efficiently solved through HMS Core Huawei's flexible audience segmentation function, combined with Huawei push, A/B testing, remote configuration, in-app messaging and other services!
Flexible audience segment creation
With Huawei Analytics Service, you can flexibly create audiences in three ways:
① Freely define audiences based on behavioral events and user attributes.
User events: refers to the various behaviors of users when using the product, including the way the user interacts with the product.
Such as login account, game level upgrade, click in-app message, add to shopping cart, in-app purchase, etc.
User attributes: refers to the basic information and status that users will not easily change.
Such as city, device type, application version number, hobbies, etc.
The combination of user attributes and user events can completely describe "who did what at what time".
For example, through the audience analysis function of Huawei Analytics, you can define “users who have purchased in-app purchases more than 3 times among old Huawei mobile phone users in the past 14 days”, “new users who have not logged in in the past 3 days”, and “not renewed after expiration. Eastern China users with membership fees" etc.
② Existing audiences get together, merge and differ, and combine them into new audiences
When creating an audience, choose the "audience combination" method, such as excluding churn users from all users, and a new audience is generated: non-churn users.
Or for the three existing audiences: Huawei Mate 40 users, male users, and users older than 30 years old, a new audience is generated: Huawei Mate 40 male users younger than 30 years old.
③ Intelligently create audiences through analysis models
In addition to the above two ways of autonomously creating audiences, you can also generate audiences with one click from the funnel, retention, and life cycle models analyzed by Huawei.
For example, in the funnel analysis report, you can choose to store the incoming and outgoing users in a certain part of the funnel as audiences with one click;
In the retention analysis report, click the number of people in the table to save the next day retention and 7th retention users in a certain day or a certain period of time as an audience.
In the user life cycle report, novice period, growth period, mature period, silent period, users at each stage, high-risk users, and high-potential users can be stored as audiences.
How to apply audience segments
1. Insight into the behavior and attributes of the audience and provide a basis for hierarchical operations
For segmented audiences, you can view its event distribution, system version distribution, model distribution, location distribution characteristics, such as analyzing "users who have paid more than 1,000 yuan in the last 14 days" and "the cumulative paid amount in the last 14 days is less than 1,000 Yuan users", whether there are significant differences in their event distribution and model distribution, etc.
In addition, you can also study its behavior characteristics in detail in other analysis reports.
For example, for the audience of "new members added in the past 30 days", you can filter the audience through filters in the path analysis report to view the characteristics of their behavior paths; you can also view the distribution of the audience’s startup period in the startup analysis report. View their favorite pages in the page analysis report.
Through the audience segmentation function, you can also divide users into core users, active users, inactive users, and lost users according to the frequency of use of core functions; or divide users into first-tier city users, second-tier city users, third-tier city users, etc. , Carry out targeted user hierarchical operation.
For example, operational activities that focus on payment promotion can focus on covering core users, while users with low potential value and inactive users are more difficult to convert, and a plan needs to be developed to gradually advance; operational activities such as promotion can target silent users. Consider the way of profit stimulus, and for new users, do a good job of novice guidance and newcomer gift package activities.
2. Assist accurate delivery and fine operation
Audience segmentation is also very useful in accurately pulling new scenes. For example, according to business needs, high sticky users are stored as audiences, combined with Huawei's rich analysis report, and multi-dimensional in-depth observation of the behavior characteristics and attribute characteristics of high sticky users, such as customer acquisition channels, age groups, and core functions of high sticky users. Use frequency, behavior path characteristics, etc., so as to provide data support for the formulation of new directions and strategies.
Secondly, audiences created through Huawei analysis can be directly synchronized to Huawei Push, A/B testing, remote configuration, in-app messaging and other services to help refine operations.
-Combine Push for precise reach: For example, send coupons to receive messages for high-risk users predicted by the user life cycle model, and send Push for lost users in the funnel payment link.
-Combining A/B testing and choosing the best solution: Audiences created through Huawei analysis can be applied to A/B testing, such as testing the application interface, copywriting, product features, or marketing activities for different groups of people. User feedback is used to make plan selection and improve the accuracy of decision-making.
-Combining in-app messages to increase the conversion rate of active users: You can create audiences through Huawei analysis and send in-app messages to active users. For example, push purchase reminders to users who have joined the shopping cart but have not placed an order.
-Combining remote configuration to accurately customize application content or style for different users: For example, provide tailor-made content, appearance and style for users of different genders and hobbies. Or trigger an upgrade reminder in multiple scenarios for users of the old version.
The above is a brief introduction of Huawei's audience analysis model to promote lean operation of mobile applications in multiple scenarios.
After the SDK of Huawei Analytics Services is integrated, the developer can obtain the user attributes and user behavior data of the application with the user’s authorization and consent. These two types of data can be aggregated to study which users did what at what time to form data The basis of analysis. On the basis of these atomic data, the analysis service provides a wealth of analysis models to help developers gain insights into user growth, user behavior characteristics, and product functions, and can perform drill-down analysis through the filter function to trigger further operations action. It is worth mentioning that the HMS Core Analysis Service SDK supports all terminals, including Android, iOS, and Web, and it only takes half a day for developers to successfully integrate and release.
For more information about Huawei analysis, please visit the official website of "Huawei Developer Alliance">Development>Analysis Services to find us quickly. We look forward to your joining!
Original link:https://developer.huawei.com/consumer/cn/forum/topic/0201458311699230845?fid=18&pid=0301458311699230633
Author: HUAWEI-HMS Core