Making good use of user feedback--talking about the processing of user feedback data

For many colleagues engaged in product work, "user-centric" is the focus of the work, but also difficult. The user's mind is elusive, and the work of user research seems inscrutable. However, user research does not always have to be done using eye-movement testing as a professional tool. With the help of many products now have "user feedback" function, in fact, can be a simple user data analysis.

"User feedback" as a means of user research, has a unique advantage. First, "User feedback" reflects the problems that users are taking in real-world use. Focus groups, user interviews and other methods, in predicting the user's behavior habits can play a huge role, but lack of user actual operation of the data, usability testing can provide data on user behavior, but the test environment is very different from the user's actual use environment, so the conclusion is still predictive. "User Feedback" goes a step further and effectively collects the problems encountered by users in the process of using the product. Second, the implementation of the "User Feedback" feature is simple. It can be a button on the software interface, or it can be a "FAQ" page on the website, or even a hotline and a person to answer. Greatly reduced the threshold of user research.

The role of user feedback analysis

Before analyzing "user feedback", make a clear analysis of what "user feedback" can do and cannot do (as shown below). Otherwise it is easy to let analysis flow on the surface of the text, or by the user a wide variety of ideas led by the nose, can not hit the nature of product problems. Through the analysis of user feedback can let us do three things: first, learning the user's language, from the user's perspective to understand the product, so as to build the user's mental model of the product. Second, understand what the user's expectations are, what expectations are met in the product, and which are not met. Third, understand the user in the use of the product "pain point", that is, the most disturbing the user to use the product of the problem.

Also note that the two points are: first, "User feedback" collection of recommendations does not represent the feelings of all users. Because even if the "user feedback" mechanism threshold is very low, it will make a lot of less enthusiastic users in the face of the problem to remain silent. Second, the direct use of the user's words to describe the problems of the product may be a risk. Be aware that users may have been annoyed by the problems they were having when they made their comments, or that they simply confused the use of the product and could not calmly and accurately tell you what the real problem is.

Analysis Method of user feedback

The user feedback analysis, can make the user fragmented output, lack of organizational information systematization, easy to quickly find product problems. The analysis process of user feedback is divided into three steps: coding-> The feedback classification by encoding-> analysis of the classification results (as shown below). The following is a brief introduction.

1, coding

In general, the coding and analysis of user feedback only extracts data from the last few months to be sufficient to illustrate the problem. In this time range from the "user feedback" system to extract the feedback, according to the content of the feedback set up coding, until no new coding. Coding can be any dimension, as long as it helps to follow up on the analysis, such as coding (performance issues, interaction issues, new functionality expectations, etc.), or coding according to the user's feelings (anger, disappointment, satisfaction, exceeding expectations). Note that you need to focus on the facts when reading feedback and coding, not because the tone of user feedback is so intense that the problem is serious.

Next, we need to extract a certain amount of feedback from the recent user feedback (100-200), which is classified according to the code by two people respectively. When finished, the two men compared the results of their respective classifications. Verify that the two people agree on the same code? Is there an encoding that does not make sense for product improvements? Is there a code that is too general or too narrow? is the actual content the same, but naming different encodings? Based on the investigation of the above problems, the coding system can be fine-tuned, so that other analysis of the feedback is clearly classified.

2, classification

After coding is established, the feedback items that are not involved in the coding process in the recent feedback data are classified according to the defined coding system. It is important to note that, because of the existence of "silent" users, the number of feedback bars in each coding category is difficult to reflect the importance of the problem and cannot be easily concluded.

3, analysis

When analyzing the classification encoding data of user feedback, we should first pay attention to the additional information of the data. For example, the data comes from a user group of what it is, what time the data is collected, what work the user is doing with the product when generating feedback, and so on. Secondly, we should observe the comparison and change of the data. For example, what is the difference between feedback from two user groups, how the user's feedback content changes at different times, and so on. It can be imagined that if there is a significant increase in negative feedback after the release of a new version of a product, the changes in the relevant parties will likely impair the user experience.

Because space is limited, the analysis method of user feedback is summed up here first. In the long run, the systematic classification and analysis of user feedback can help us to predict the direction of future development of our products by keeping abreast of the changing trends of users ' ideas and the response of users to new versions of products. Moreover, the user feedback analysis has the characteristics of easy data and low professional requirements of the analyst, which is very suitable for the simple User Data analysis tool.

References: Kuniavsky, M. (2003). Observing the user Experience:a Practitioner ' s Guide to user. San Francisco,ca:morgan Kaufmann.