As the demographic dividend of the Internet continues to decline, it is increasingly important to retain the existing users of the product. And Internet access to user feedback and product was more traditional usage way, when a user Internet products use bug and had to buy a car in the use of feedback process and found problems in the same time, will be looking for customer service, customer service with robot will play with you first, and then is the problem record, a long wait, don’t rush back, go away. Internet products are different from traditional industrial products, and the replacement cost of users is not much, except for a few national level apps.

Automobile has a hundred years of development history, so far there is no guarantee that every car production will not have problems. What's more, Internet products have not had a bug after just one or two decades of development. But there is nothing "Internet" about the way we deal with users' problems, and it takes a long process. Now, the way to obtain information has changed from users' active search a few years ago to feed information flow and short video recommendation today. Then, can we also be smart in quality assurance and service quality, so that users' anger when they encounter problems with products can be changed to actively find problems and provide solutions? Let every problem into a product active touch users of the opportunity, also let users really feel the warmth of the product. From the perspective of quality assurance, under the background of rapid iteration of Internet products today, the test engineers are greatly challenged. First of all, the new release speed of domestic mobile phones brings great challenges to compatibility testing. Second, the product function and technical framework are becoming more and more complex, and the evaluation scope of each iteration is becoming larger and larger. Therefore, it is unlikely to be completed in a short time to test all aspects of regression. Although many products have adopted the grayscale release to reduce the impact of bugs on users, no matter in grayscale or in full scale, if problems occur to a small number of users or in full scale, the impact is still real, these users still have the possibility to choose to leave.Copy the code

Based on this, the author boldly proposed a testing direction —- user quality portrait. Different from the popular user portrait, which collects data from users’ material and behavioral attributes, quality portrait is more about data collection when users encounter product use problems. If the release of a product feature is “usable” by users, then the quality of the product or feature is “usable” or “durable” by users. How to establish user quality portrait the author thinks there are several aspects as follows:

Collect user data in case of problem

The data collection is mainly divided into two parts: one is to collect the data actively, and establish a unified quality center at the client end, front end and back end of the APP, so that network problems, bugs or other problems can be determined through the data collection. The second is passive collection, which collects problems reported by users in user feedback, screenshot sharing and other behaviors.Copy the code

Analyze user quality data

After collecting user quality data actively and passively, be able to analyze the cause of the problem. This is a certain adjustment for the current Internet test engineers. Data analysis ability was previously considered as a professional ability, and any Internet practitioner should have professional data analysis ability in the future. There is still a long way to go from the data to the essence of the problem. It is impossible to find the essence of the problem simply by analyzing the data, and sometimes the opposite conclusion can be obtained.Copy the code

Build channels to help users

When we find a problem and can fix it, how do we let the user know? Now when problems are found in the product, they are quietly fixed, and the version update log is drowned in meaningless text like “fixed some bugs”. At this time, you need a channel to touch users, no matter the app push, or the way of short message, have not been used in quality problems. Imagine if we fixed a problem and sent users a message that said, “You have had some problems with our product in the past. Today, it is online and optimized. Every time you use our product, we have not forgotten your problems.” When a user receives a message like this, whether or not he is using the product, he will want to see if the problem is really solved.

The construction of user quality portrait is not limited to the above aspects, and the specific practice of each part will be written in detail later. Readers who are interested in this are also welcome to discuss together.

This article is published by OpenWrite, a blogging tool platform