User Segmentation for Multi-Scenario Precise Operations - Huawei Developers

Products must fulfill wide-ranging user preferences and requirements. To enhance user retention, it is important to design targeted strategies to achieve precise operations and satisfy varying demands for different users. User segmentation is the most common method of achieving this and does so by placing users with the same or similar characteristics in terms of user attributes or behavior into a user segment. In this way, operations personnel can formulate differentiated operations strategies targeted at users in each segment to improve user retention and conversion.
Application Scenarios​In app operations, we often encounter the following problems:
1. The overall user retention rate is decreasing. How do I find out which users I'm losing?
2. Some users claim coupons or bonus points every day but do not use them. How can I identify these users and prompt them to use the bonuses as soon as possible?
3. How do I segment users by location, device model, age, or consumption level?
4. How do I trigger scenario-specific messages based on user behavior and interests?
5. Can I prompt users using older versions of my app to update the app without having to release a new version?
...
The audience creation function of Analytics Kit together with other services like Push Kit, A/B Testing, Remote Configuration, and App Messaging helps address these issues.
Flexibly Create an Audience​With Analytics Kit, you can flexibly create an audience in three ways:
1. Define audiences based on behavior events and user labels.
User events refer to user behavior when users use a product, including how they interact with the product.
Examples include signing in with an account, leveling up in a game, tapping an in-app message, adding a product to the shopping cart, and performing in-app purchases.
User labels describe user attributes and preferences, such as consumption behavior, device attributes, user locations, activity, and payment.
User events and labels allow you to know which users are doing what at a specific point in time.
Examples of audiences you can create include Huawei phone users who have made more than three in-app purchases in the last 14 days, new users who have not signed in to your app in the last three days, and users who have not renewed their membership.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
2. Create audiences through the intersection, union, or difference of existing audiences.
Let's look at an example. If you set Create audience by to Audience, and exclude churned users from all users, then a new audience containing only non-churned users will be generated.
Here is another example. On the basis of three existing audiences – HUAWEI Mate 40 users, male users, and users whose ages are greater than 30 – you can create an audience containing only male users who use HUAWEI Mate 40 and are younger than 30.
3. Create audiences intelligently by using analysis models.
In addition to the preceding two methods, you can also generate an audience with just a click using the funnel analysis, retention analysis, and user lifecycle models of Analytics Kit.
For example, in a funnel analysis report under the Explore menu, you can save users who flow in and out of the funnel in a certain process as an audience with one click.
In a retention analysis report, you can click the number of users on a specific day to save, for example, day-1 or day-7 retained users, as an audience.
A user lifecycle report allows you to save all users, high-risk users, or high-potential users at each phase, such as the beginner, growing, mature, or inactive phase, as an audience.
How to Apply Audiences​1. Analyze audience behavior and attribute characteristics to facilitate precise operations.
More specifically, you can compare the distributions of events, system versions, device models, and locations of different audiences. For example, you can analyze whether users who paid more than US$1000 in the last 14 days differ significantly from those who paid less than US$1000 in the last 14 days in terms of their behavior events and device models.
Also, you can use other analysis reports to dive deeper into audience behavior characteristics.
For example, a filter is available in the path analysis report that can be used to search for an audience consisting of new users in the last 30 days and view the characteristics of their behavior paths. Similarly, you can check the launch analysis report to track the time segments when users from this audience launch an app, as well as view their favorite pages, through the page analysis report.
With user segmentation, you can classify users into core, active, inactive, and churned users based on their frequency of using core functions, or classify them by location into users who live in first-, second-, and third-tier cities to provide a basis for targeted and differentiated operations.
For example, to increase the number of paying users, you are advised to focus your operations on core users because it is relatively difficult to convert inactive and low-potential users. By contrast, to stimulate user activity, you are advised to provide incentives for inactive users, and offer guidance and gift packs to new users.
2. User segmentation also makes targeted advertising and precise operations easier.
User segmentation is an excellent tool for precisely attracting new users. For example, you can save loyal users as an audience and, using a wide range of analysis reports provided by Analytics Kit, you can analyze the behavior and attributes of these users from multiple dimensions, such as how the users were acquired, their ages, frequency of using core functions, and behavior path characteristics, helping you determine how to attract more users.
In addition, other services such as Push Kit, A/B Testing, Remote Configuration, and App Messaging can be used in conjunction with audiences created via Analytics Kit, facilitating precise operations. Let's take a look at some examples.
Push Kit allows you to reach target users precisely. For instance, you can send push notifications about coupons to users who are more likely to churn according to predictions made by the user lifecycle model, and send push notifications to users who have churned in the payment phase.
Applicable to the audiences created via Analytics Kit, A/B Testing helps you discover which changes to the app UI, text, functions, or marketing activities best satisfy the requirements of different audiences. You can then apply the best solution for each audience.
As for App Messaging, it contributes to improving active users' payment conversion rate. You can create an audience of active users through Analytics Kit, and then send in-app messages to these users. For example, you can send notifications to users who have added products to the shopping cart but have not paid.
What about Remote Configuration? With this service, you can tailor app content, appearances, and styles for users depending on their attributes, such as genders and interests, or prompt users using an earlier app version to update to the latest version.
That concludes our look at the audience analysis model of Analytics Kit, as well as the role it plays in promoting precise operations.
Once you have integrated the Analytics SDK, you can gain access to user attributes and behavior data after obtaining user consent, to figure out what users do in different time segments. Analytics Kit also provides a wide selection of analysis models, helping paint a picture of user growth, behavior characteristics, and how product functions are used. What's more, the filters enable you to perform targeted operations with the support of drill-down analysis. It is worth mentioning that the Analytics SDK supports various platforms, including Android, iOS, and web, and you can complete integration and release your app in just half a day.
Sounds tempting, right? To learn more, check out:
Official website of Analytics Kit
Development documents for Android, iOS, web, quick apps, HarmonyOS, WeChat mini-programs, and quick games