Teach you to build four systems of user operation and three classic growth models.

What is the user operating system? I believe that every enterprise has a relatively perfect user operation system. Before I came into contact with some articles introducing user system, I basically equated user operating system with user grouping strategy and AARRR operating mode. In fact, this is only the tip of the whole operating system. Combined with the experience in the operation process, I explored a set of practical user operation system in the process of user operation, including four strategic systems: growth framework: user growth team+core growth channel+growth tools. User modeling: user model construction, including label portrait, user value model, user preference identification model, user churn early warning model, user activity model, etc. Scenario layering: 12 operation scenarios, each scenario is layered and grouped based on user tags and modeling tools, and corresponding precise marketing means are formulated for operation. Data operation: core operation index system+data analysis model. First, the construction of the user growth framework is self-evident. Without user growth, user operation is impossible. The common practices are: the marketing department operates the recruitment channels and is responsible for the delivery and optimization of various application malls and online advertising platforms; New media recruits new media operations and is responsible for the content output of social channels; The user team is responsible for activating, promoting and retaining users in the user pool. At first glance, this team system is completely fine, each responsible for one piece and each achieving KPI. In the actual operation process, the following inevitable problems may occur: the boundaries of departments and the unity of KPI settings are different. Based on the above problems, the primary task for most enterprises to do user growth is to establish a user growth team. The growth team must first eliminate departmental boundaries and exist as a project team or growth department, including channel operation, activity operation, product operation and user operation. Secondly, each operation node based on AARRR defines the growth index for each function to guide the whole growth work: the main assessment of channel operation of acquisition node: new users, acquisition cost (CAC) and retention rate of new users. The main assessment of products at activation and retention nodes: registration conversion rate and functional retention rate; DAU, gross, DAU/ gross users in revenue and reference nodes; As an independent growth team, user conversion rate and K factor only assess one department indicator at the KPI level, and each function is linked to this indicator, thus solving the problems of fragmentation and mutual shirking. We found that some products that are alive and better must have their own core growth channels. Mobike's body QR code has gained enough riding users through offline delivery, Didi's red envelope has gained enough taxi users through sharing channels, and Pinduoduo has gained enough e-commerce users through social channels. My social crowdfunding platform also relies on a strong offline traffic portal and online fission mechanism to form a strong closed-loop product channel in the creation of core growth channels, and quickly realize user growth and retention. Whether it is an independent growth team or a departmental collaborative growth team, it is necessary to learn a variety of growth tools to assist themselves, improve work efficiency, better judge growth links and adjust growth strategies. Construction of User Growth Framework As a growing person, we should first make a 90-day growth plan; second, build a user model. If an enterprise can't even build a basic label portrait model, user operation can only be an armchair strategist. The construction of user model is the basis of realizing hierarchical grouping of users, and it is also a necessary tool for users to operate accurately. User models include label portrait model, user value model, user preference identification model, user loss early warning model, activity model and so on. The value of labels lies in helping operators to achieve scene stratification for users based on their services and design targeted marketing activities. The value of portraits lies in helping operators understand the characteristics of each group; User value model can identify high-value user groups; Preference identification model helps operators to push products in a targeted manner; The early warning model of churn can retain users before churn, and the activity model can wake up and promote activities in a targeted way. The establishment of the model needs a special data product team to complete. When marketing based on user model, operators need to pay attention to marketing effect analysis and iterative optimization of marketing scheme. Through many marketing attempts and data product teams, we found a more suitable modeling method and gradually established a stable operation plan and operation plan. When going to work every day, operators can combine the labels used up the previous day into marketing information of the user group and send it out (push or SMS), supervise its transformation, iterate constantly, and gradually establish a standard operation plan and operation plan based on the user model. Third, the strategy of scenario stratification is based on platform business, and several operation scenarios can be derived. Each scenario needs to operate different user groups, and the user groups come from the tag model and each user model. In the specific operation process, our operation is divided into two categories: one is growthhack, and the other is refined operation by users. Iv. user data operation strategy data operation includes core index system and data analysis system. The core index system can monitor the development trend of users' operation and know the basic information such as user activity and health in real time. User data analysis system can help operators locate problems in time and optimize products. The first is the construction of core index system, which must be closely combined with product objectives. At the same time, people at different levels in the enterprise pay different attention to the core index data, and the leadership level pays attention to the volume, cost and income of large-scale users; The operational level focuses on user activity, retention and transformation; In the construction of index system products, we construct the core indicators of consumer users from four dimensions: new customer acquisition ability, health, preference and purchase behavior. 1. Analysis of the growth potential of new customers: cities, stores and promoters understand the overall development and development potential of users in regions, business districts and communities; User source channel analysis: every channel wants to know the current push channels, and which channels do users mainly come from? Which channels are of high quality, so as to optimize the channel strategy; Analysis of new products: stores and places want to know which products in the region contribute the most to new products, and the products that customers place orders for the first time are defined as new products; Analysis of novelty-seeking preferences in various communities: Stores and promoters want to know the preferences of new users in various communities in the region, such as: A community prefers electronic products and B community prefers fresh food, so as to promote novelty-seeking in various communities. 2. User Health User Value Analysis: The channel wants to know who its loyal users are and can find these high-quality users to let them participate in the activities. Similarly, promoters can invite these users to the store to participate in activities online; User churn index: the channel wants to know which different groups of users will be churned out and how to prevent them from being churned out; Community users' contribution: stores and promoters want to know the GMV contribution rate of each district where the promoters are located, and the distribution in the district should have weekly and monthly trend charts. 3. User preference category preference: stores, promoters and channels want to know which community/region is more inclined to consume what kind of goods (cross relationship between buyer's location and category); Activity preference: stores, promoters and channels want to know which community/region prefers what kind of activities (the cross relationship between the buyer's location and activities); Price preference: the channel wants to know what prices users of different categories prefer, so as to push various price segments to the corresponding users (the cross relationship between categories and prices); Contact preference: stores, promoters and channels all want to know which channels users of different categories prefer to buy (the cross relationship between category and contact). 4. Repurchase rate of different user groups: The channel wants to know the repurchase rate of new and old users and find out high repurchase products, adjust the operation strategies of new and old users in time and do a good job in commodity operation, and monitor them on a monthly basis. User path analysis: the channel wants to know the user participation from the channel home page to the active page, and in which link the user is lost, so as to do a good job in page operation. Secondly, the data analysis system needs to build a series of analysis model tools to help operators locate the problems in the operation process. Model tools include funnel analysis model, attribution analysis model, micro-transformation analysis model and queue analysis model. Take the low registration conversion rate as an example to briefly describe the analysis method: step one: disassemble the influence dimension; Step 2: Disassemble the subdivision indicators under the dimension; Step 3: Locate the problem. Registration conversion rate can be divided into two dimensions: channel and product. Decompose the segmentation indicators under each dimension, and the channel segmentation indicators include media delivery, advertising types, advertising content and keywords; Products include registration logic, product design, input method, product stability, etc. To locate the problem, we need to check the subdivision indicators one by one, find the abnormal data points, check the conversion rate of each link through the funnel, and focus on the links with low conversion rate. If it is a channel problem, optimize the media, AB test the advertising content, and accurately locate the keywords; If it is a product problem, optimize the registration logic and interface to improve the stability of the APP. Summarizing the four strategic systems, we can find that user operation is no longer a simple job of finding a few operators to do a good job of group operation, nor is it a job that a few user models can quickly enhance the user value of enterprises, but an operation system that enterprises need to invest manpower, energy and material resources for a long time. The significance of user operation to enterprises is self-evident. The growth of the overall performance of an enterprise is inseparable from the expansion of the scale of high-quality users and the improvement of the life cycle value of users. Three classic growth models 1, user operation target |AARRR model What is user operation? It aims at maximizing user value, and improves activity, retention rate or other payment indicators through various operational means. In different life stages of products, according to the depth and types of users' participation behavior, users' operational objectives can be divided into five parts, namely, acquisition (acquiring users), activation (activating users/improving users' activity), retention (improving retention), income (increasing income) and recommendation (spreading viruses), which is the so-called AARRR model. Acquisition This is the first step in operating the product. In this step, we need to answer two questions: Who is my user? How to get it? Who is my user? You need to describe the user portrait according to the product attributes, including core functions, business logic, product stages and so on. , find the scene where the target users gather. How to get the target user after finding it? This is actually a channel problem. At present, there are various channels to acquire users, including seed users, social media communication, search engine optimization, offline promotion and so on. , and the specific way and how to choose the focus, according to the different types of products and different stages of development. Activation (enhancing user activity) For most products, what really matters is the number of active users rather than the number of registered users. Because only active users are creating value for products, we need to improve product value by increasing user activity. At this stage, two questions need to be answered: What is the user's product experience? What kind of benefits do users get? Because only a good product experience and predictable benefits (material rewards and emotional rewards) will keep users active in this product. Retention (improve retention) "users come quickly and go quickly." This is an inevitable phenomenon brought about by the popularity of the Internet. We can see many phenomenal products, such as Traveling Frog, Face Meng and Footnote, which were very popular at first and gained the user's ability MAX, but often only lasted for a short time. If you want to improve user retention, then you must constantly increase the abandonment cost of users, that is, increase the contact and interaction between our users and products, and guide them to do more output/payment on our products. On the contrary, the low retention rate of some products is related to the product form, such as take-away products and O2O door-to-door service. Users will hardly care too much about whether such products provide personalized needs, and the only thing they care about is cost performance. In the era of internet revenue, almost all products are for profit, and the essence of how to make a profit is the establishment of product business model. Some products are born with a stable business model, but there are also many tool products that have not formed a business model, such as mailboxes, players, picture reading software and so on. Their purpose is often to improve the company's product line, form a closed loop, and enhance the user's experience of the product. The product itself doesn't actually make money. In addition, it can be said that many companies are slowly exploring a feasible business model after the products are almost formed. For example, sports apps such as Yuepao Circle, Glug and Keep are basically "advertising+e-commerce (sports equipment)", and a large part of Yuepao Circle's profits come from online marathons. For example, Get, Himalayan FM and other audio products, through the launch of "paid products", have changed from an audio sharing platform to a knowledge payment application platform, opening the road to commercial realization. Referral virus transmission, also known as self-transmission, refers to the explosive growth of users based on word-of-mouth promotion such as sharing and interaction in online and offline social interaction. The rise of many brands is promoted by word of mouth, such as the "three squirrels" of nuts. Considering that customers need garbage bags when eating nuts, a bag is set in the package. This meticulous and thoughtful experience quickly established word of mouth among customers, and customers voluntarily shared this wonderful experience on social platforms, further establishing online word of mouth, thus achieving explosive growth of users. Of course, not all products can spread like viruses. The key lies in whether we understand users' psychology, whether we can meet users' needs, and whether we can bring users unexpected experiences, so as to arouse users' desire to share and make Internet products achieve the maximum communication effect at the lowest cost. By fully understanding and applying AARRR model, we can clearly understand the operational objectives of products in different life stages and formulate different operational strategies accordingly. 2. User Hierarchical Management | User Pyramid Model In front of us, we used AARRR model to analyze the users as a whole, and briefly outlined the different problems to be solved in different operation stages. But in fact, with the increase of user base, users begin to show different attributes (such as gender, region, age, etc.). ). Even users with the same attribute have different product behavior habits, and the user group is no longer a simple whole. As operators, we can't adopt a "one size fits all" simple and rude operation mode, and we need to operate according to different groups of people. This is the so-called refinement strategy and user stratification. 1. The first layer is community managers, usually product operators; 2. The second layer is user management tools, such as moderator system and community (interest groups, tribes, guilds, etc. ); 3. The third layer is valuable users, such as VIP and KOL；; 4. The fourth level is the general users, that is, the "80% users" in the 28 th principle. Those products that have been used in the community can basically be disassembled with this model, the most typical one is Xiaomi; In Xiaomi Forum, the top of the pyramid includes not only Xiaomi's operators, but also developers, designers and engineers. They do quick iterations every week based on user feedback. However, relying solely on Xiaomi's official team is far from enough to manage this forum with tens of millions of users, so Xiaomi helps users and manages users by developing moderators, advisory groups, cool play groups, city clubs and other communities among users. The third-and fourth-level users of the forum basically follow the "28" principle: 20% users have produced 80% of the content of the community (even lower), and it is these 20% users who have promoted the establishment of community culture and affected those 80% users. Of course, the "user pyramid model" is not completely applicable to all products, because not all products have their own communities, but the operating logic behind it can be applied to other product forms. For example, the user stratification of e-commerce can do this (the user's use path is simplified to "registration-use-order-payment-sharing"): after the user stratification, the user can be targeted. For example, we want new users to start using products, and the common strategy is newcomer welfare; We hope interested users can complete the decision to pay and buy goods. Common strategies are promotion (through limited time discount) and so on. To sum up, the inspiration of the user pyramid model to operators is: 1. Through user stratification, targeted operation strategies are adopted for users with different characteristics; 2. Spend 80% of users' operation time with the top 20% users of the pyramid, because they are the most valuable core users; Only those who have been in contact with users for a long time can understand the real needs of users. 3. User Grouping Management |RFM Model The user hierarchy is an up-and-down structure, but this structure cannot completely summarize the user groups. Earlier, we marked the paying user layer as "paying", but this group is also different. Some users spend a lot of money, some users buy it frequently, and some users have bought it but don't buy it now. How should this be subdivided? If we continue to increase the number of floors, the conditions will become complicated, which is not conducive to the formulation and implementation of operational strategies. Therefore, we adopt the horizontal structure of user grouping and continue to subdivide the grouping at the same level to meet the needs of higher precision and refinement. RFM model is a classic method of customer management. It is a typical clustering model to measure the value and profitability of consumers. It depends on three core indicators of charging: Recency, Frequency and Monetary. Last consumption time: measure the loss of users. The closer the consumption time is to the current users, the easier it is to maintain the relationship with them. The value of users who spent 1 year ago is definitely not as good as that of users who spent 1 month ago. Consumption amount: measure users' contribution to enterprise profits. The higher the consumption amount, the higher the value. Consumption frequency: a measure of users' loyalty is the number of purchases made by users in a limited time, and the users who buy the most have higher loyalty. For example, important value customers: high consumption amount and frequency, and close consumption time in the near future. This is a core user and needs maintenance. It is important to retain customers: the frequency and amount of consumption are high, but the recent consumption time is far away, indicating that this is a loyal customer who has just arrived. You can actively keep in touch with him by SMS, push or EDM. From the foregoing, it can be seen that the RFM model constructs a consumption model through three dimensions: the latest consumption, consumption frequency and consumption amount, which is mainly suitable for products with multiple repurchase requirements. However, if we are dealing with Weibo, Today's Headlines, Tik Tok and other products, payment is no longer an indicator to divide user groups. In this case, we can replace the three dimensions of R, F and M with three new dimensions, namely, the last opening, the number of times of opening and the duration of use, to build a user activity model. You can use RFM model to establish different grouping strategies, and more gameplay is waiting for everyone to explore together. Today's summary: the construction of growth system includes four strategic systems: growth framework: user growth team+core growth channel+growth tools. User modeling: user model construction, including label portrait, user value model, user preference identification model, user churn early warning model, user activity model, etc. Scenario layering: 12 operation scenarios, each scenario is layered and grouped based on user tags and modeling tools, and corresponding precise marketing means are formulated for operation. Data operation: