Users are lost, and it is difficult to reach?
The campaign has done so much and the conversions are still low?
Operations also need to think ahead, predict user trends, and reach out accurately in advance, in order to achieve more with less. Combined with the prediction service and intelligent operation of HMS Core analysis service , it can gain insight into marketing opportunities, implement marketing strategies in real time, and improve user operation efficiency.
The forecasting service has accurate forecasting models and in-depth crowd insights, supports viewing the forecast accuracy of the past week, and helps operators make more scientific operation strategies.
Intelligent operations provide periodic scheduled tasks and real-time trigger tasks to help operators reduce operating costs and enable continuous and accurate activity conversion.
Pre-churn cycle recall to improve in-app retention
Re-engaging churned users is difficult, so predicting the risk of churn ahead of time is critical.
For example, when users have the risk of product experience loss or competition loss, they can predict in advance through user-related feature data, segment users who may be lost in the next week, and plan and carry out corresponding pre-recall activities for this part of the audience. users, which can effectively improve user retention.
Usually operators also try to predict churn based on existing in-app data or their own experience, and take some measures to delay churn, but the accuracy rate may not be high.
The prediction service uses AI algorithms, combined with machine learning technology, to establish and train relevant prediction models to help developers discover potential target audiences from audiences and greatly improve prediction accuracy.
*How the forecast service works
For example, game applications generally pay more attention to user activity, set periodic tasks for high-probability churn populations through intelligent operations, and send in-game gift activities when users are close to churn, so as to improve the effectiveness of user retention and improve user activity.
* Pre-churn user recall
Real-time access to key nodes to promote user conversion
The monetization ability of a product is an important consideration for the healthy development of a product.
Many times, we find that the product has high activity, but the final paid conversion rate is not ideal. Paid prediction accurately subdivides users who may make purchases in the next week, and automatically creates audience groups with three probabilities: high, medium, and low. You can also customize the interval probability according to the actual operation needs of the product to predict the groups that match the actual characteristics of the product.
For this group of people, operators can formulate targeted marketing strategies or optimize the payment process, and then combine the real-time marketing capabilities of intelligent operations to push activities when users reach a certain payment node, thereby increasing the user purchase rate and promoting the benign operation of application business.
*The figure is dummy data
Taking photography and beautification tools as an example, for high-probability paying users, a limited-time discount for members is offered when users use props, which can reduce the psychological game in the payment process based on the user's original willingness to pay, thereby promoting its application. In-app purchase.
*Intelligent operation - real-time marketing flow chart
In addition, intelligent operations support multiple access channels such as Push, SMS, and Webhook, and support applications to fulfill business growth demands such as accurate access; support A/B testing and activity effect analysis, effectively conduct prediction verification, and adjust operation strategies in real time.
Learn more details>>
Visit the official website of Huawei Developer Alliance
Get development guidance documents
Huawei Mobile Services Open Source Warehouse Address: GitHub , Gitee
Follow us to know the latest technical information of HMS Core for the first time~
**粗体** _斜体_ [链接](http://example.com) `代码` - 列表 > 引用
。你还可以使用@
来通知其他用户。