Comprehensive analysis of user operations

The preface recently browsed the recruitment information of Internet operation, from which I felt that the position of operation was subdivided and systematized, and the demand for user operation also showed convergence, so I could finally give it a complete definition. I hope that graduates who are curious about user operations, or graduates who have just entered the industry or are still looking for a job, can help us learn more about this work. If you find it useful, please like it+collect it ~ I need your help to upgrade my account ~ ~ 1. Description of the needs of large manufacturers for user operation posts (September 2020) In order to clarify user operation, I dug up the needs of 26 user operation posts from Tencent, JD.COM, Netease and Byte Recruitment official website, and then got their * * * information from three directions: tools/applications, objectives and capabilities. As can be seen from the above data, in terms of "tools/content", the most important thing is "activities" (such as activity operation and activity planning, etc. ), followed by "stratification" (that is, different users use different operating strategies), and the following, such as growth (system), user life cycle, research, user portrait, user motivation, etc. It seems that it is less mentioned, but it is actually doing activities. In terms of goals, activity and retention are more prominent, followed by innovation, which basically covers all the goals in the user's life cycle. In terms of ability, in addition to the general requirements such as pressure resistance and cross-departmental communication, data analysis is the focus (I believe that operation itself is also a data-driven work), and only one of them explicitly mentions SQL plus points. Second, what does the user operation do and the relationship between each transaction? How are many terms mentioned in the first part reflected in the whole operation? The following figure can intuitively understand this summary. Generally speaking, user operation is based on user life cycle and user growth model. Work aimed at innovation, promotion, transformation (first purchase, repurchase, key behavior transformation) and retention. As for the strategy or means to achieve the work goal, it depends on the demand of specific products. Usually, an app that meets the needs of routine reuse, such as food, clothing, housing and transportation, works as shown in the above figure: users are stratified, and then the stratified users are investigated to get portraits, and users with different portraits are reached in different ways according to the user's purpose and needs (usually taking activities as the landing carrier), and then the data viewing effect is analyzed and iterated. In order to let you know more about these relationships, I have drawn a picture (roughly understood as a plant [product] that can produce different flowers and an evolved bee [user]: (emm, I'll just take a look, and skip here if the relationship is clear). Looking up from the root, each branch is hierarchical, and the different flowers growing on different branches are active or surface interaction, and the fragrance of flowers is touching. Let users (alien bees) discover products and interact with them (collecting honey/spreading pollen). Different flower shapes/smells attract different bees, and different flowers determine the interaction mode of bees. The left branch is an incentive system. Many flowers are planted, and the difficulty of collecting honey from each flower is different. Bees accumulate evolutionary points by collecting pollen from primary flowers, and after evolution, they can collect pollen from advanced flowers. As for what flowers look like? Why can attract a certain kind of bees is the user survey and user portrait completed in the early stage of stratification. After many tests, the flowers finally grow into what bees (users) like. (Of course, users' honey-picking behavior can further feed back their acceptance of product performance. Third, the focus is divided into detailed goals and basic ideas: 1, goals. There will be a certain gap in the specific analysis and fineness requirements of these terms in different businesses. For example, most apps refer to new registrations, and some shopping apps refer to new users. For example, many apps will be retained the next day and retained in the third month; However, for low-frequency tool apps, it may tend to remain on the 7 th. This will be observed in detail in the specific work. 2. Life cycle is a generalized model, including the life cycle of users and the life cycle of products. Of course, the main thing here is the user. In the initial stage of acquiring users, we may do channel launch or fission activities, which will consume a certain cost, so the user value is negative at first; With more and more interactions between users and products, the value of users begins to rise (different products are different, some may be paid by users, and some may be advertising value); When the user sinks or loses, if the product is still maintaining the user, such as SMS, email, etc. , and the user does not respond to the relevant key behaviors, the user value here will drop to a negative value. There are many articles about the life cycle value of users on the Internet, so you can spend some time if you are interested. Work content part: What is stratification? Broadly speaking, stratification is a way of thinking of refined operation, which is actually to classify users in a certain way. If divided by life cycle, it means new users, active users, lost users and so on; If user value is divided into important value users, important development users, important retained users and general value users, why should it be layered? From the perspective of user development, different users have different use needs and payment habits, and their familiarity with app is also different. Differentiating operation for different users, providing different product consultation and services will reduce users' search cost, improve user experience and increase users' stickiness and conversion rate. From the perspective of operational efficiency, operational manpower is limited. If users are classified and stratified, users who contribute more to the target value will be given more operation time and higher operation efficiency. How to stratify? Need to extract data according to the goal/guess, analyze the data, and get different user groups; Through the method of investigation/active contact, the conjecture/goal is verified and the stratification is further optimized. (Sometimes stratification is not achieved overnight, and it needs repeated testing and verification to get the best stratification standard. The more detailed, the more layered standards, such as "the purchasing group of maternal and child products among important value users". There are two limiting criteria. Common layering method: 1, as shown in the above figure (layering according to work objectives is enough) 2. Value stratification: RFM How to stratify users and realize refined operation? Using RFM User Value Model (Chirs) is a very detailed article. It needs to be added that the principle of this method can be extrapolated to many aspects; And we can pay attention to its main mathematical logic: take the user proportion and observation index as YX axis, delimit the observation index according to the curve change, and then carry out the next assignment operation. (It doesn't have to be RFM, even if you define other standards yourself, users can be layered in a similar way. ) 3. Element weights are layered. After stratified user portrait survey (knowing who you are and what you like, user portraits are generally divided into several categories: 1, demographic portraits: age, gender, occupation, education, income, geographical location 2, user behavior portraits: browsing habits (browsing time, browsing duration, browsing category), buying habits (average purchase order, purchase category, purchase frequency, correlation. Lifestyle: Demographic data can be obtained through simple questionnaire survey or user ID card information, while behavioral data can be collected directly from system data. User attitude data can only be obtained through long-term interaction between users and products, which usually refers to the following data: satisfaction, perceived competitive position and position, desire characteristics, lifestyle, brand preference, social and personal values, concepts and preferences. Demographic data and preferences: behavioral data portrait (part) * schedule, user preference data and their corresponding values When understanding user preferences, we often need to know the following data: actual behavior, rating and market segmentation (segmentation is continuous, and segmentation can be gradually strengthened from the initial RFM; Methods include profitability segmentation, demographic segmentation, channel utilization segmentation, RFM segmentation, personal attitude segmentation, preference segmentation), frank dialogue (indicating preferences, answering questions, putting forward needs), and third-party information (including demographic data, resumes, living habits, etc.). ) usefulness: a high value b medium value c low value (this list is to remind yourself and your classmates that when we spend a lot of time on something/data every day, we should occasionally stand up and look at the relationship between what we do and the actual task value. In the last article, we mentioned the pre-launch operation of the event. You can look through it, so I won't go into details here. & ltahref = ""> The essence of "Deep Anatomy of Activity Operation (I): Pre-activity Preparation (including detailed flow chart and a lot of details)" is basically the above two pictures. It should be noted that 1 For user operations, every activity is a tool to understand users. Regardless of the success or failure of the activity, as long as users interact with products through the activity, we can learn more about users. Therefore, whether the activity is organized or not by the user operation, the user operation can pay more attention. (What activities did the user participate in? What does he pay attention to the way of activities and rewards? How active was he in the past and during his activities? What activities are not enthusiastic) 2. In the complete process of stratification, activity can be used as the last link to verify the conjecture. The user incentive system is still the previous article "Advanced Operation: Analysis of the Bottom Logic of User Incentive System". The key points of the full text are as follows: User incentive system is actually the practice of guiding user behavior and enhancing user value with the peripheral value of products. To be an incentive system, the premise needs to be clear. What is the value of a user's behavior? * Some products blindly enter the points system integral mall and other functions. Only users can get points by keeping (daily login/sign-in/browsing) and simply accumulating some time to exchange physical gifts, which is easy to attract bonus hunter users and increase operating costs. So what factors should be considered before establishing an incentive system? How to mine users' value behavior? How do user value and calculation score match equally? Please see the previous long article: "Long Raiders of User Incentive System: From Bottom Logic Analysis to System Establishment Description" IV. Daily work of user operation (for reference only). Then we directly look at the daily work of user operations (only for the refinement of their own work experience, not necessarily universal). In fact, the work of operation is basically the same, and it also belongs to the work of multi-departmental communication. For new operations, the key is to quickly get familiar with background applications, product processes, division of labor and working methods of various departments/groups, communication processes, etc. In addition, because the user operation carries the goals of activity, retention and transformation. They may also undertake touch planning (SMS, email, WeChat, push), resource planning, coupon distribution planning and so on. Because it's not in the hot words of recruitment, I won't go into details. * More than mentioned above, the purpose of user operation is to acquire valuable users, strive to retain users and improve user value. Therefore, senior operations can go in this direction, combine the company's goals, and then try new work content and projects, and don't limit their creativity to a few known terms. V. Enlightenment of CRM on User Operation (Thinking) The analysis of CRM is quite consistent with user operation, from which we can also get some enlightenment. 1, target interpretation Give up unprofitable users: In the description of CRM target, "Give up unprofitable customers" is also part of the target. In the daily work of user operation, there is a kind of "lost user recall", which usually uses SMS, ITU and so on. , complex and low conversion rate, requiring labor costs and high communication costs (domestic SMS costs are low, but some). Reducing service costs and operating costs is the key point of growth. So when we measure the value of our work, we can also add cost reduction, such as automatic response system to reduce the cost of manual communication. 2. Management mode IDIC agrees with the unification of multi-channel and multi-system. When customers come, it quickly identifies the type of customers and treats them differently from any important interactive information in the past. Leave time for high-value users to interact efficiently. Every interaction is based on the previous one, and it undertakes the customer-centered personalized service of the previous interaction (which requires the trust of customers and more two-way communication). It should be noted that this step requires more technical support. Because it needs to integrate multi-channel information, for example, the user feeds back questions through applets and communicates with customer service, then when he later moves to the app, the system also needs to identify the questions that this person feeds back in applets and the related communication progress. (The biggest contradiction in this step should be the security of private data. Differentiated treatment is also the key, including reserving time for high-value users, and also requiring differentiated treatment for users according to the differences in customer needs. 3. Customer value (profit contribution) The customer's contribution to the enterprise's profit is 1, and the profit is obtained by increasing purchases; 2. Reduce operating costs and gain profits (customers are more experienced, demand extraction times are reduced, and error probability is reduced); 3. recommendation; 4. Price Overflow Long-term customers are more likely to pay normal prices. 6. Conclusion This paper mainly focuses on the concepts and thinking logic involved in user operation. Originally, I wanted to add some examples of operation, but I always thought it was too complicated, so I had to put it behind. The content of the article is mainly based on personal experience and thinking. If you have other ideas, please leave a message.