Tencent's user growth methodology and practice
Seizing and sharing Tencent’s user growth methodology and project practices is something that Sanbai has always wanted to do . It took me nearly a month and more than 30 nights of overtime to finally finish writing it today . I urgently want to share it with everyone and invite friends who are also struggling on the road of business growth to communicate and discuss with me. In order to help you better understand the context of the full text, I will organize the entire content into a mind map as follows: As a reminder, all the data involved in this article are from public data, and some case data are examples that require non-real data. 1. Write in front1. Why do I want to share this content?(1) Develop your own user growth methodology As an Internet veteran with 8 years of experience, Sanbai has worked in first-tier Internet companies such as Tencent, Kingsoft Office, and Vipshop, and has always focused on the field of business growth. He can be regarded as a practitioner who has worked deeply in the field of business growth. I worked in the years around 2015 when e-commerce was booming and personalized recommendations were prevalent. Through PUSH personalized recommendations on e-commerce platforms, the click-through rate of the daily PUSH push notifications reached hundreds of millions, increasing by 70%, contributing tens of millions of yuan in daily revenue. When domestic office software and SaaS were popular, I witnessed more than 70% year-on-year growth in annual office subscription revenue and paying users. In 2020, when video conferencing came onto the domestic stage during the epidemic, I experienced the growth of a video conferencing product from nearly 0 to a user base of hundreds of millions. These three experiences of business growth have enabled me to understand and learn a wealth of Internet-leading business growth-related knowledge and methods in specific enterprise project practices. Especially when I was at Tencent, I really wanted to sort out and consolidate these methodologies to form my own set of business growth methodologies. (2) Help friends who are also exploring user growth In August 2022, I began to think about whether these methods and knowledge could be replicated to startups with limited resources to help them find the direction of growth. So this year, I provided user growth consulting services to friends from three startups, helping them to sort out the direction of user growth and follow up on their strategy implementation. During this experience, I felt that many small and medium-sized enterprises and entrepreneurial groups generally had the following problems: 1. No clear direction for user growth I have seen many small and medium-sized enterprises that do not even have a user growth strategy or plan for their own products. The product, operation, sales and other teams are all doing their own things, and no one knows where to focus the company's resources. 2. Lack of scientific decision-making Even if some companies have a clear direction, the decision-making method for this direction is made by a few people after brainstorming. No one tries to verify it, and then they start to pass it down and let the executive level do it. 3. Wrong way of doing things and verifying, leading to wrong conclusions Some companies have come up with a strategic direction through a series of logics, but when it comes to implementation, due to incorrect methods of doing things and verification, they use some result data that may not stand the test, draw some wrong conclusions, mislead everyone, and cause everyone to quickly abandon this idea; I often hear some old employees in the business say that every time someone proposes an idea, they will say that we have tried this idea and the conclusion is that it doesn’t work, and they have done many things but they have not been verified to work. However, when they take out the original data and take a look, they find that the verification method is not eye-catching at all. Based on these insights and experiences of helping others, I would like to write down these problem-solving methods, hoping to help more people. Maybe my current level of cognition and knowledge is not enough, but I will do my best and keep improving. 2. Who is it suitable for?(1) Entrepreneurs/Startups It is suitable for friends who already have an existing product or service, a clear business model, and are thinking about how to achieve user growth. Here we assume that your product or service is mature and the business model is clear, because this is the premise of user growth. We will not discuss the business model and other pre-issues in this sharing, otherwise these will be a big proposition when discussed separately. Let's focus on one point: the product or service is already there, how to achieve user growth? (2) Managers of the user growth team Team managers who are in charge of the company's user growth business, those who want to output a user growth plan, and corporate workers who want to learn and understand user growth methodology. (3) Internet product managers/operators Friends who are product managers and operations personnel in Internet companies and want to find some models to solve practical problems. 3. What content should be shared?Regarding the user growth series, I have prepared sharing on 5 topics, including the 6-step method for user growth, how to build a user growth team, data indicator systems and tools, commonly used products and operation tools, and how to disrupt innovation and find the second growth curve; the core content is the 6-step method for user growth, and this article will only include this part, and I plan to output the rest in succession. Regarding the 6-step method for user growth, the overall sharing content will include the following two parts:
4. What will you gain?1. Master Tencent’s user growth methodology, and explore and practice the direction and strategy of user growth for your product like Tencent’s people do. User growth is a science! Because my team members at Tencent came from various products and businesses of the company, and in the process of working with them, I found that the growth ideas and methodologies of most of Tencent's businesses are basically similar, so if you really understand the content shared in this article, you can really explore and practice user growth like Tencent people; 2. Learn to output a logically rigorous and persuasive user growth strategy plan In this article, I will provide specific steps and tools for exploring user growth strategies, including data analysis, user research, thinking models, etc. Using these tools, you can output a user growth strategy that can stand the test of time, is logically rigorous, and convincing. 3. Learn a series of user growth models to solve practical problems I believe that in your past learning experience, you will more or less understand various growth-related models, such as addiction model, user experience map and many other messy models. What Sanbai hates is that these models are very scattered and each one solves a single point of problem. I have a strong desire to connect these models together and put them into the entire thinking chain of exploring growth, so that each can play a role. In this sharing, I will introduce the problems you may encounter in the practice process in a full-link manner, as well as the reference models for solving specific problems; 5. Correct some misconceptionsAnother problem is that many people think that achieving growth in Tencent is an easy thing. For example, many of my friends would say, isn’t it very easy for Tencent to achieve growth? Just divert traffic to traffic entrances such as WeChat and QQ, and the growth will occur. It sounds like that, but the reality is that the product managers within Tencent are worrying about growth targets every day. On the one hand, having internal traffic support is indeed an advantage, but it is only an advantage for cold start. For example, if you want to achieve 100,000 or even 1 million MAU in the short term, it may be no problem to attract traffic, but how to retain traffic after attraction is a problem. Tencent will not be satisfied with just one million MAU. Usually we want tens of millions or hundreds of millions of scale, which is impossible to achieve by internal traffic. In Tencent's history, many products finally grew, and the result was not just traffic. On the other hand, the traffic of WeChat and QQ cannot be attracted just because you want to. It is harder than climbing to the sky to put an entrance in their products, especially WeChat. If your introduction does not create value, Tencent's product managers will not agree to such traffic cooperation. 2. What is user growth?When we talk about growth, many people always think of spending money on advertising and other practices. Sorry, these are not our definition of user growth. In this article, we are more discussing how to drive user growth in a "free" way. I like to quote Qu Hui, the author of "Silicon Valley Growth Hacker's Practical Notes", to define user growth:
From Qu Hui’s definition, we can see that the study of user growth has the following characteristics:
3. How did Tencent Meeting achieve user growth from 0 to 100 million?1. The product with the fastest user growth in Tencent’s history!As a star product of Tencent that was born during the epidemic, Tencent Meeting has grown from less than tens of thousands of monthly active users to over 100 million registered users in less than a year, even exceeding the speed of WeChat. According to public data, as of September 2023, the number of registered users of Tencent Meeting has reached over 400 million. Even though the epidemic is over, it still maintains a relatively high active scale and user retention, and is still the most used product in the domestic video conferencing track. 2. Isn’t the growth of Tencent Meeting due to the epidemic and the various traffic diversions from Tencent?When it comes to the growth of Tencent Conference, many people will take it for granted that it is due to the epidemic and Tencent's inherent traffic advantage. As a person who has experienced it personally, looking back on this issue after a few years, I think this way: It is true that the rapid growth of Tencent Conference in the early stage was definitely due to the stimulus of the epidemic and the investment of resources by Tencent. Referring to the change curve of the Baidu Index of "Tencent Conference", it can be seen that the stimulus brought by the epidemic came and went quickly. After a short period of rapid growth in early 2020, it soon began an unstoppable cliff-like decline. From September 2020 to the end of 2021, after the number of active users dropped to a low point, its active users began to maintain a sustained and stable growth. The impact of the epidemic during this period has been very small, and the marketing budget of Tencent's conference business has been significantly reduced during this period. Its growth relies more on product innovation and growth strategies. It must be admitted that the epidemic has indeed caused a concentrated outbreak of market demand for video conferencing, which has shortened Tencent Meeting's growth path by at least 3 to 5 years. However, even without the epidemic, the domestic video conferencing market and demand still exist (this has been confirmed by Zoom in the US market), it just arrived a few years later. Therefore, my opinion on the growth of Tencent Meeting is: it started with the epidemic but is not limited to the epidemic, was born in Tencent but is not just because of Tencent. 3. What has Tencent Meeting done to drive user growth?So, what exactly did the Tencent Conference team do to drive the continued growth of its product? Growth is the result of multiple factors. To be honest, it is difficult to attribute it to just a few things. However, I think Tencent Meeting has at least done the following things right to drive its user growth:
Of course, each of the four points mentioned above is actually very large and abstract, so I will mainly list two successful growth cases and share how we gradually determined and verified the strategy through the 6-step user growth method:
Key conclusion: The growth of Tencent Conference started with the epidemic but is not limited to the epidemic; it was born in Tencent but is not only because of Tencent; the growth of Tencent Conference benefited from four points: open platform, user-first product concept and ultimate product experience, expanding competitive product differences and continuous expansion of product ecology. （1）Identify the North Star Metric for User Growth Many businesses use DAU, WAU, and MAU to measure the scale of user growth. In order to quantify the user growth target, we use the "monthly number of users attending meetings" as the North Star indicator of user growth, that is, the number of users who attend meetings within a month. The logic of this design is as follows: Why choose conference attendance as the criterion for activity instead of launching an APP? Because we believe that the purpose of users using Tencent Meeting is to attend meetings, so effective activity means that users want to attend meetings. Simply starting it is not effective activity. Why monthly statistics? Video conferencing is not a high-frequency user demand. Users do not use video conferencing products for meetings every day. Sometimes even large meetings are only held once a month. Therefore, in order to measure growth relatively reasonably, we choose to count on a monthly basis. Therefore, the "monthly number of attending users" is what we define as MAU, which is the North Star indicator of user growth that we define. Key conclusion: The number of monthly conference participants is the North Star indicator for Tencent Conference user growth (2) Identify the appropriate user growth driving model ①What did we do at the beginning? When it comes to user growth for a video conferencing product, in the beginning, no one knew where to start. Our growth team has many colleagues who have experience in user growth for C-end products such as QQ, so at the beginning, based on their work experience in C-end operations, we tried to carry out operational activities such as attracting new users, providing user welfare incentives, and even branding cooperation with super apps within the company such as Peace Elite, trying to increase the scale of conference users through C-end operations. ② Wake up from ineffective attempts! As you can imagine, these actions did not produce much growth effect. We were frustrated to find that whether users hold meetings or not is determined by demand. Users will not come to meetings just to receive a reward or participate in an operational activity. What we did did not play a big role. For video conferencing, it is a product whose growth is driven by product functions rather than operations. Therefore, in the process of gradual exploration, we gave up the operation mode that was mainly based on operation gameplay and focused more on the product's usage value and usage scenarios. We later verified that this change was relatively correct. ③Understand the user growth driving model of the product to avoid going in the wrong direction from the beginning This experience also taught us that different types of products have different growth driving models. Wrong growth driving is like adding diesel to a gasoline car. At the beginning of designing a user growth strategy, we should first figure out the core factors that drive user growth, otherwise it is easy to work in the wrong direction. In the second part of the method and model section, I have summarized 4 common growth driving models for Internet products. I will not elaborate on them here for the time being. Key point: Recognize the user growth driving model of the product to avoid going in the wrong direction from the beginning. Tencent Conference is a product driven by product growth. (3) Identify the core levers for user growth After all the messing around, we started to think about the direction of Tencent Conference's core growth strategy. In other words, if we only focus on a few strategies this year and concentrate resources on doing just a few things, we can achieve that difficult user growth goal. I call this the core lever of user growth (in layman's terms, it is a strategic direction with lower investment costs and better growth effects), and how do we reasonably determine the core lever? To sum up, we mainly used the user growth formula model to gradually identify the core levers of user growth. This model can be summarized into three steps: Step 1: Define the user growth formula for the product and identify the key user roles for growth Choosing a suitable user growth formula For the user growth formula of the MAU (monthly number of users attending the conference) indicator, we considered two calculation methods: Growth formula 1: MAU = MAU of conference hosts + MAU of conference participants Growth formula 2: MAU = conference host MAU x average growth coefficient
At the beginning, our logic was based on the first growth formula. Under the guidance of this formula, the goal of increasing the number of meeting users was to increase the scale of both host users and participant users. Under the guidance of this strategy, all our actions were equally focused on the growth of these two groups. For example, in advertising, we placed ads to both potential host groups (such as teachers, HR, etc.) and participant groups (such as students, corporate employees). After operating for a period of time, we found that this approach would have the following problems: a. This is a strategy that evenly distributes resources, covers all aspects, and has no focus or focal point; b. In a meeting, the host is the key decision maker and influencer. The products he uses determine what products the participants use. c. At unit cost, the contribution of a new host to growth is N times that of a new participant; Based on these findings, we began to realize that we should focus our resources and attention on the role of the host. The core goal of user growth in the conference scenario is to increase the scale of host users. The growth of hosts will drive the growth of participants. So we changed the logic of the user growth formula and changed to using growth formula 2. In other words, the core goals of MAU growth are two: a. Increase the number of conference host users; b. Improve the average growth coefficient of hosts; Step 2: Drill down and analyze the growth formula to confirm the direction of the growth strategy After confirming the growth formula, the next step is to further analyze the growth formula and determine the strategic direction of core growth based on different dimensions. For the host MAU indicator, we mainly drill down based on the user activity status and crowd dimensions, and for the average fission coefficient, we mainly drill down based on the conference scenario, as follows: ① Host MAU indicator: Based on the user activity status, confirm the priority of the new, retained, and return strategies. We divide the scale of host MAU into three parts according to the user activity status: new hosts, retained hosts, and return hosts, among which: a. Newly added hosts: refers to newly added host users in the current month; b. Retained hosts: refers to host users who were active last month and continue to be active this month; c. Returning hosts: refers to hosts who were active in the past, but inactive last month, and active this month; This is a more mainstream user segmentation method, which can clearly show the user stratification status of active users and know the active types of active users; Therefore, from the perspective of user activity, there are only three strategic directions for host user growth:
Learning from the previous lessons, we always remember one thing: don’t do everything, but focus on one thing. So which of these three points should we prioritize? Our thinking at the time was as follows: a. From the perspective of growth value: We have accumulated hundreds of millions of registered users. With the same investment, retaining an active old host user can bring N participants; while adding a new host user can only bring M participants (N>M); therefore, the value of improving the retention rate of old host users is higher than that of adding new host users. b. From the perspective of difficulty: the final conversion rate of acquiring a new user from trigger to conversion is X%, while the probability of a lost old user returning is only Y% (X>Y); so from the perspective of difficulty, returning is a relatively difficult task, and the priority of adding a new host is higher than the return of the host. Moderator MAU indicator: further breakdown based on target groups to identify key target groups We have understood the importance of retention and new additions. The next step is to know which users should we focus on retaining? Which users should we add in terms of new growth? We can further break down the target population. Usually, just looking at the distribution structure of the population may not be very meaningful, because we need to know which core population should be retained or added. Therefore, we can confirm the core target population by building a four-quadrant model of user value: a. Construct a four-quadrant model of the penetration rate of different groups of retained hosts + the host's next month retention rate In this part, we mainly confirm which target group should be the core focus among the retained hosts, among which: the host penetration rate of a certain group = MAU of hosts in the target group / MAU of hosts in the overall market, which represents the scale value of the target group. The host retention rate of a certain group of people in the next month = the number of users who hosted meetings last month and are still hosting meetings this month / the number of users who hosted meetings last month, which represents the target group; the four-quadrant model of different user groups established according to these two indicators is as follows (the content and data in the figure below are purely fictitious, not real data, and are only used as examples) : In this model, we should focus on the groups in these two quadrants:
b. Construct a four-quadrant model of potential growth scale + fission coefficient for different groups of new hosts In this part, we mainly confirm the angle of adding new hosts, and focus on which target groups should be added, including: The potential growth scale of a certain group of people = the scale of active users in the entire market of that group - the scale of active users in my product, which represents how much growth room I still have in the entire market. The fission coefficient indicates how many participants a new host user can bring, representing the fission value of customer acquisition. According to these two indicators, we also establish a four-quadrant model for different user groups as follows (the content and data in the following figure are purely fictitious, not real data, and are only used as examples): In this model, the quadrant we need to focus on is the population in the high-scale and high-fission quadrant. He said that this part of users still has a relatively high growth space, and the fission value of acquiring this part of users is higher. We can see from the figure that teachers, HR/administration, and KOL groups are our first priority target user groups to acquire. (3) Average fission coefficient indicator: breakdown based on different meeting scenarios For the indicator of average fission coefficient, we further break it down according to different meeting scenarios to look at the fission coefficients of different meeting scenarios. For example, let's take a few example scenario data: the fission coefficient of the online teaching meeting scenario may reach 40 to 50 people, the fission coefficient of the corporate training scenario can reach 20 to 30 people, and the fission coefficient of the interview scenario is usually 2 people, because most interviews are 1V1 interviews. Therefore, in order to improve the overall fission coefficient, we need to increase the penetration rate of meeting scenarios with high fission coefficients. In order to determine which meeting scenarios’ penetration rates need to be improved, we also establish a four-quadrant model. Among them, the conference penetration rate = the number of meetings in this conference scenario / the overall number of meetings in the market. In this model, we need to focus on the scenarios in the quadrant of low penetration and high fission. These scenarios have higher fission value, but the penetration rate is relatively low, so we need to improve the penetration of these conference scenarios. From the figure below, we can see that the scenarios of corporate training and online teaching are conference scenarios that need to focus on improving penetration (the following data is purely fictitious, not real data, and is only an example). Step 3: Prioritize growth strategies and identify core growth levers using the ICE scoring system The ICE scoring system is a model mentioned by Sean in Growth Hacker. It is used to evaluate how to determine the highest priority strategy model when you have many growth strategies and ideas. The model mainly prioritizes various options from three dimensions: influence, confidence, and simplicity. We try to use this model to prioritize the previous strategic directions (the following information is fictitious and is only used to illustrate the problem): Combined with the results of the ICE score, we can see that in the growth strategy direction mentioned above, our highest priority is to find ways to increase the retention rate of teacher users and the user scale of the new teacher group, while increasing the conference penetration in online teaching scenarios and corporate training scenarios. This is what we define as the core growth lever for user growth. Key points summary: We gradually identified the core levers for user growth in three steps
The entire logical deduction process and conclusion above are summarized in a picture as follows: 4. Find the magic number for user growth(1) Why do we need magic numbers? In step 3, we confirmed the core lever for user growth, but this lever strategy is usually a strategic direction, or it is generally a conclusion given to the boss level, but this conclusion cannot be used directly to guide execution. When we communicated the above conclusion to the executors of various business parties, we found that everyone understood the theory, but still didn’t know how to do it. It felt that the thinking was still not so clear because the strategic direction did not form quantitative indicators that could be implemented. So the next step is to transform the goal of achieving this strategic direction into the goal of executable items, and it would be best to have a quantitative indicator. We call this the magic number of user growth. (2) How to find the magic number? The following is mainly about how we analyze the magic number of achieving the core leverage target through the three steps of "user research → logical assumptions and ranking of the importance of influencing factors → correlation analysis". Let's take how to improve the retention rate of teachers as an example: Step 1: Build awareness based on user research To improve the retention rate of teacher users, we first need to know why teacher users are willing to retain and why they are unwilling to retain? After the survey, we learned that the main reasons why teachers are willing to stay include:
The reasons why users are unwilling to stay are usually as follows:
Step 2: Logical assumptions and ranking of influencing factors Based on the previous user survey, we can preliminarily assume that the factors affecting the retention rate of teacher users may include the following:
So far, the above factors are just assumptions. Whether these factors really affect the retention of teacher users and to what extent, we evaluate their impact through data modeling analysis. We will ignore the specific modeling here, which involves more data analysis and engineering knowledge. Let's look at the results directly (the following results are fictitious examples, not real data, and are only used as examples): According to the analysis results, the factor that has the greatest impact on retention rate is frequency of use, followed by the number of functions used and the number of usage scenarios. Whether or not to use the screen sharing and cloud recording functions has little impact on the retention rate , so it may not be a factor affecting retention. Step 3: Correlation Analysis In the second step, we know that the frequency of use, the number of functions used, and the number of usage scenarios are factors that affect the retention of teacher users. So when these indicators reach a certain level, the retention rate of teachers can be significantly improved. We establish a correlation analysis between these factors and the retention rate, and draw a relationship curve between the retention rate and each influencing factor (reference example is as follows). From the analysis results, we can see:
"Host N meetings in 1 month", "Use M functions", "Hold Y meetings in different meeting scenarios". These numbers are the magic numbers we need to find to achieve our goals. Under the guidance of magic numbers, we understand that to improve the retention rate of the teacher group, what we need to do is to find a way to get the teacher group to host at least N meetings within one month, or use M product functions, or hold two scenario types of meetings (such as class, class meeting, etc.); this is the clearer and more executable than the previous strategic strategy. For students at the execution level, they also know what to do. This is the importance of magic numbers. It points us to the executable quantitative goals for achieving results. There are many cases in the industry about magic numbers. We will give additional explanations when reviewing the second part. Key points: Magical numbers that improve teacher user retention rate Preside over N meetings within 1 month"; 2. "Use at least M functions"; 3. "Hold at least Y meetings in different meeting scenarios"; Similarly, I tried to summarize the above content in a picture as follows: 5. Design growth strategies based on magical numbersMagic numbers allow students at the execution level to find the direction of executable implementation. Next, we will focus on implementing the quantitative indicators of magic numbers and talk about how to design growth strategies and what common ideas and thinking models can be used to learn from. Similarly, we take the goal of "let teachers and users host N meetings within one month" as an example. Our thinking process is as follows: (1) Let users complete the golden experience path of the product once, and accelerate the experience of Aha moments Let’s first popularize the concept. The so-called golden experience path represents the operation path that needs to be completed in the process of obtaining the key services provided by the product. Aha moment refers to the links in the entire experience path that make you feel happy and refreshed. The obvious truth is that if you want to find a way to get the teacher to host N meetings within one month, you must first let the user host a meeting completely and complete the product experience process of the hosting meeting. Otherwise, let alone let the user continue to use N times in the future, if the teacher does not complete this golden experience process, it will be difficult for the teacher to become a host user who can be stably retained. So how should we let users complete the experience path, and how should we speed up the AHA moment on the experience path? Within Tencent, we often use two models: user experience map and emotion curve model, and "conversion = desire-friction" model. Model 1: User Experience Map and Emotion Curve Everyone is familiar with this model, so I don’t need to introduce it too much. The main functions of the user experience map and emotional curve include 2 points:
Next, we try to draw the user experience map and emotional curve of the teacher group's use of Tencent meetings. The use of conference products usually includes three stages: pre-meeting meeting stage, meeting start stage, and post-meeting stage; the three stages constitute the user's complete user experience, so we look at the user experience map and emotional curve of each stage respectively. 1. Pre-meeting meeting stage - User experience map and mood curve From the emotional curve, we can see that there are 2 emotional troughs and 2 peak experience points in the scheduled session of the entire meeting: (1) Emotional low point a. Select the meeting type link: The reason why users experience emotional troughs in this link is that they don’t know much about the differences between the three meeting types: regular meetings, special meetings, and webinars. Therefore, in order to help users understand, we may need to add some clearer and more vivid annotations, such as telling users what meeting scenarios are suitable for each meeting type. b. Configuration meeting information stage: The reason why users have a low mood in this link is that the configuration information list is too long and some configuration items are not known for what they are used for. Therefore, in this link, we may need to reduce some infrequently used configuration information, and only most of the configuration items needed by users are exposed by default, while some configuration items required by subdivided scenarios are hidden through collection. (2) Peak experience point a. Click on the scheduled meeting session: The reason why this session has peak experience is that the homepage of Tencent meeting only has 3 buttons by default. Users can quickly find the entrance when entering the homepage, which greatly reduces the difficulty of users finding the entrance. This has been maintained since its launch. Although the homepage has been discussed many times, this point has not been changed. b. Complete the reservation link: When the user sees that the meeting reservation has finally been completed and the reservation success page is seen, the appointment will naturally be aroused. How can we find a way to enlarge this peak experience? We have imagined that the function of generating conference promotional posters is added in this link, because sharing pictures when sharing meetings will be more fulfilling than sharing text. In addition, we also imagine that users who book the meeting for the first time can also give resources rewards. 2. Conference Beginning Phase - User Experience Map and Emotion Curve Similarly, when we look at the beginning of the meeting, we can see 2 mood troughs and 3 peak experience points: (1) Emotional low point a. Turn on or off the microphone link: The main reason for the emotional trough in this link is that many people forget to turn off the microphone and then die of the society; some people also forget to turn on the microphone after talking for a long time; so around this link, we have imagined some optimization measures. On the one hand, many people hope that the microphone should be in the default state when joining the club, so the microphone should be turned off by default (later we adopted this method, and the microphone should be turned off automatically when the number of participants exceeds 6). In addition, in order to avoid the situation where the microphone was not turned on after talking for a long time, we have also added an automatic detection function. If you detect that you are talking but the microphone is not turned on, you will remind you to turn on the microphone. b. Waiting for students to join the meeting: This kind of scenario is very common. Usually, meetings with multiple people often require waiting for at least 5 to 15 minutes before they can get together. Therefore, in order to avoid this situation, we imagine that the function of participants collecting meetings and reminding before the meeting should be strengthened; (2) Peak experience point a. Pre-meeting reminder: When receiving pre-meeting reminder, the teacher and user will feel happy because they avoid forgetting themselves; for amplifying this peak experience, we intend to find ways to increase the reach channels and reach rate, so we have added the function of binding calendar and calendar reminder. b. Turn on the camera: The reason for the peak experience in this link is that after turning on the camera, you can see your avatar with a beauty effect. The beauty function is very popular among many people; for amplifying the peak experience, we envision it to add richer filter functions and add virtual headgear and other gameplay to make this link more fun and interesting. c. Check-in: Tencent Meeting has built-in check-in application, which efficiently solves the problem of check-in efficiency of teachers and users and is deeply loved by teachers, so later we have added more meeting applications one after another. 3. Meeting End Phase - User Experience Map and Emotional Curve At the end of the meeting, we can see more peak experience points. On the one hand, after the teacher finishes the meeting, he can view the meeting statistics and conference recording, which is very useful to the teacher. On the other hand, we also support the function of users who can redeem their own meeting duration for rewards, which makes users' emotions reach a climax, because the meeting duration is too long, which is annoying, but if you can redeem the rewards, you will feel great comfort. Model 2: Transformation = Desire - Friction This model comes from Sean's book "Growth Hacker". The core logic is that if you want to improve users' conversion of the golden experience path, the core strategy is either to increase users' desire to take a complete path, or to reduce users' friction on this path. In the Tencent Meeting product, let's think about these two issues: 1. What are the frictions on the path of gold experience? It is usually difficult to enhance desire, and reducing friction is usually relatively simple, so let’s first think about what product optimizations we can do around the idea of reducing friction: a. Shorten the golden experience path: For example, we have made an optimization point that can support users to join the meeting directly without logging in. This greatly shortens the process of attending the meeting and has also won praise from many users; for example, we also support the function of clicking on the meeting number to directly open the meeting details page and merge the meeting. There is no longer a need to copy the meeting number to open the Tencent Conference APP to re-enter the meeting. b. Reduce friction in key loss links: For example, during the meeting, the teacher group will encounter noise in the surrounding environment, or delays and lags in the bad network environment, resulting in a very bad user experience. Therefore, we have spent a lot of energy on the ability to reduce voice noise, and the ability to maintain a high fluency in a weak network environment. 2. How to enhance users' desire to complete the golden experience path How can we motivate users to continue to go down the experience path without interrupting and giving up, and be willing to keep going, usually our thinking is: a. Set stage achievements on the path: For example, the design of the meeting details page for the scheduled meeting stage mentioned above is a better case; b. Reducing the cost of using is also a way to enhance motivation, because users will feel that the cost is low anyway, so it doesn’t matter if they try it. For example, the homepage of Tencent meeting mentioned above has only 3 buttons; Model 1 and Model 2 actually represent the idea of thinking about growth strategy design. You can actually see that these two models actually contain a lot of similar logic and have similar effects. We just provide two ideas for design strategies. You can choose which model to use according to your understanding and preferences, or you can also use the two together. Remember that our purpose is to design strategies, and models are just tools. (2) Use addictive model to allow users to continue using The main thing we thought about before was how to solve the problem of users completing a golden path. Next, we need to think about how to make users continue to use it. In this link, we use the HOOK model (also called the addiction model) to think about this problem and organizational strategy; the HOOK model is mainly used to sort out and explain why users continue to use a product and service. We often use it to understand the reasons for users' continuous retention, and at the same time, we also think about growth strategies to improve user retention based on these reasons. Next, let’s try to use the HOOK model to understand why those teachers who are willing to host more than N meetings in a month are willing to continue using Tencent meetings? 1. Use HOOK model to analyze the reasons for retention of Tencent meeting users a. Triggered: The main reason why most teachers use Tencent meetings is that they have the need to teach at home due to the impact of the epidemic. Then, they use Tencent meetings for official recommendations of the school, recommendations of teachers around them, recommendations of school service providers, and marketing advertisements. This is an external trigger; then, because most teachers have used it once, they feel that it is quite useful, and can solve the problems of online teaching and conference communication well. I feel the benefits of using Tencent meetings, so I am willing to continue to use it. This is an internal trigger. b. Generate actions : The reason why many teachers use Tencent meetings without DingTalk is also because they understand that we provide many functions for online teaching scenarios, such as recording, online testing, naming and other functions, and participate in the experience function, which feels very suitable for their needs. They have developed experience behavior, so they retain it. c. Rewarded: Many teachers feel satisfied because students and themselves express their expressions and chat messages during online teaching. After continuous use, they will be praised by other teachers or leaders for understanding the functions of many Tencent meetings. This greatly stimulates them to continue using them; and when they see the digital records of the teacher’s courses, they feel very accomplished. d. Investment generated: The reason why the teacher has been using Tencent Meeting is because he has mastered the use of this product and has taken time to learn new tools. This is because he has already invested in the cost of learning time. Then, because the teacher’s account has accumulated a lot of course recorded data, data, etc., so the teacher will not easily change the platform. 2. Thinking about growth strategies to improve teacher user retention based on the HOOK model a. Enhance effective trigger points: Before explaining what an effective trigger point is, first explain the invalid trigger. We have done a lot of PUSH pushes that introduce how useful our products are, and then there is no big effect. Then, based on the previous research, we know that to allow teachers to continue to use Tencent Meeting, the strategies include the following points: ① The school must be handled, especially the school’s information technology department: because the teacher uses what software to teach, there are usually official recommendations and regulations for the school, and the school’s information technology department will provide tools and suggestions, so we need to overcome the school and preach our products in the school; ② Cooperate with university service providers: There are many service providers of colleges and universities, such as service providers that provide teaching systems. They do not have the ability to video conferencing. At that time, they hope to integrate video conferencing products on their platform, so we can combine the needs of service providers to place the use portal of Tencent meetings on their platform and use them to help us reach schools; b. Provide customized solutions and product functions to guide actions Provide more product functions and solutions suitable for teachers, and strengthen and guide them to experience product functions. c. Increase reward stimulus The length of the teacher's class and meeting can be used to redeem benefits such as Tencent Video VIP. The more times and duration of classes, the more rewards you can get, as mentioned above. d. Allow users to continuously accumulate assets and increase replacement costs Pay attention to the functions of guiding users to use cloud recording and other functions that are prone to asset precipitation, and give the teacher group more space for cloud recording, so that the teacher group can continuously accumulate product use asset precipitation. The more precipitation is, the less likely users to lose. Key points summary: Based on the addiction model, we propose strategic points to enable users to continue to use:
At this point, we basically obtained the design of a growth strategy to achieve magical digital goals through a series of logical deductions. Similarly, we summarize the complete content of this part with a picture: 6. Verify the effectiveness of growth strategy through AB experimentsThe previous one is about the magic number of "hosting N meetings within 1 month". We designed a series of growth strategies to find ways to let teachers and users host at least N meetings within 1 month. What we need to do next is to verify the effectiveness and effectiveness of these growth strategies through AB experiments one by one. Regarding how to do AB experiments, we will not talk about them here. In the second part, we will focus on how to do it. Among the strategies mentioned above, there are successful positive strategies and failed strategies. Next, I will show a few positive strategies we finally adopted and their effects. See below for details: 1. Positive strategy 1: Open the video conferencing capabilities to domestic education service providers by using APIs and other means, thereby increasing the frequency and retention of teacher users; Strategy content: We cooperate with top university service platforms in China, including Xuetong, Yucai, Shenzhen University Youcai, etc., to support the products of these third-party university service platforms to create meetings and arouse Tencent meetings by calling Tencent Meeting's API, helping college teachers solve the pain points of needing to use multiple products across the terminal and not having data interoperability. Strategy effect: This function has been well received by many college teachers and students, and this strategy drove the overall MAU growth at that time. 2. Positive Strategy 2: Meeting Duration Redeem Benefits Strategy content: Support teachers to use the duration of the meeting to redeem benefits such as Tencent Video Members, QQ Music Members and other benefits; Strategy effect: This strategy is also a well-received design for teachers. Many people originally complained that the meeting was too long, but when redeeming for benefits, they enjoyed the joy after hard work, and the retention rate of teachers' group was also significantly improved. At this point, the case part ends here. 4. 6-step method for user growth: method, model and conceptPreviously, I took the practical case of Tencent meeting as an example and explained in detail how to explore growth strategies through the 6-step method of user growth. In this part, we extracted some user growth methods and models that can be reused to other industries and products from the perspective of general methodology, concepts, and models. The goal is to hope that everyone can follow the finding a set of methods and processes to find the strategic direction of user growth of their products and business. Let’s review the 6 steps of user growth in the previous case:
Next, I will share in detail the specific practice and precautions for each of the steps; 1. Confirm the Polaris indicator for user growthThis step may sound quite simple, but in fact, many teams have made mistakes from this step, so it is necessary for us to pay attention to the setting of the Polaris indicator. (1) What is the Polaris indicator? The concept of the Polaris metric was originally proposed by business model creator Alex Osterwalder, whose significance lies in helping organizations focus on ensuring all activities work towards the goal of achieving the metric. This metric is often associated with the core values and strategic direct goals of the company or product, and it reflects the long-term success and sustainable growth of the organization. (2) How to design the North Star indicator for user growth? In the Internet circle, DAU and MAU are usually used as core indicators for measuring user growth, and everyone's caliber designs for DAU and MAU may also be different, but that doesn't matter. I think the design of the Polestar indicator for user growth should meet the following three characteristics: 1. Reflect active scale rather than registration scale: Usually we pay more attention to how much active is, rather than how much registration is, because registered users may have lost, and active scale can better reflect user growth. 2. Establish association with the value of the product and reflect the effective value User scale: How to define activeness? Some products will use launching APPs or opening websites as the standard for defining activeness. The problem with this method is that we may attract many invalid users through advertising and activities. This seems that the scale of active users has increased a lot, but these users are spam users and have no value generated; so I advocate more defining activeness with effective activity. For example, for a document product, effective activeness is to open or create documents; for a video conferencing product, effective activeness is to join the association; for a social product, effective activeness is to initiate IM dialogue. 3. Consider the active frequency of users and design statistical cycles reasonably: How to define statistical cycles, this usually depends on the characteristics of the product and the frequency of users' usage. For example, for an office product, users may use it once a week or once a month. The statistical cycle may be more recommended to count weekly or monthly; for products like WeChat and Douyin that you may open every day, counting by day is also possible. Summary: Designing the Polaris indicator for user growth needs to meet the following 3 points:
2. Recognize the appropriate user growth-driven model(1) Recognize the importance of the user growth-driven model In the case of Tencent meeting earlier, we mentioned that at the beginning, based on the reasons of empiricism, we tried to make a tool product with a product-based function through operational activities, user welfare stimulation and other operational methods, and the effect was almost minimal. I also cited a relatively vivid metaphor, choosing the wrong growth driver, just like a gasoline-burning car, if you add diesel to it, a product is like a car, and the growth driver model is like fuel, and then the correct one can make the growth of the product run. Since this experience, in the subsequent experience, whenever I think about the user growth strategy of a product from 0 to 1, I always think about a question first, what is the core driving model of user growth of this product? I need to avoid working hard from the beginning. (2) Common user growth-driven models for Internet products For the common user growth-driven models on the Internet, I summarized them into the following 4 types: PLG (Product-driven growth model) 1. Typical products: more typical products that drive growth, including social products represented by WeChat and QQ, as well as tool products mainly based on Tencent Meeting and WPS; 2. The core driving force of user growth: The core of user growth of this type of product mainly relies on the continuous launch of product functions that meet user needs or solve problems in a certain field of users, and continuously acquire users through continuous product iteration; 3. What to do to promote growth: So what to do to promote growth of this type of product is to discover user pain points and markets, launch better product solutions, and let users experience the pleasure brought by a good product experience; 4. What users value more: Users of this type of product will pay more attention to product interaction, design experience, product tone, etc., and they will pursue the ultimate product experience more, because the quality of these will greatly affect the user experience; OLG (operation-driven growth model) 1. Typical products: more typical OLG products include UGC content platforms such as Zhihu, Douyin, and Xiaohongshu, or e-commerce platforms such as Taobao, JD.com, and Pinduoduo; 2. The core driver of user growth: The core of user growth of this type of product mainly relies on the content, goods or services provided by the product, while the production of content usually relies on the platform's operating mechanism; 3. What to do to promote growth: What to do to promote user growth of this type of product is to provide competitive content, goods or services; for example, Douyin and Xiaohongshu, etc., ensure high-quality content output by establishing a complete KOL content production system, and for example, the staff of the Dian E-commerce platform obtain goods with price advantages through continuous stocking; 4. What users value more: users of this type of product value more the quality of the content or the price discount of the goods. As long as it is not too difficult to use, they can accept it; SLG (Sales Driven Growth Model) 1. Typical products: The more typical SLG products mainly include SAAS products, especially pure to B products, such as Sales Yi, Beisen, etc., or many business products under the Huawei system; 2. The core driving force of user growth: The core of user growth in this type of product lies in the company's sales ability. If a strong self-selling system or an agent distribution system can be established, it can ensure the continuous growth of the product; and in the initial stage, even your products do not need to be particularly strong, and the experience does not have to be particularly good, as long as it can be used; I have seen many difficult products, but because their sales system is particularly strong, they can live very well; 3. What to do to promote growth: What to do to promote growth of this type of product is very simple, that is, to continuously develop your own distributors or increase sales capabilities; 4. What users value more: For this type of product, customers value more whether the product can solve practical problems. They have higher customization needs and hope to be highly customized and scenario-based to facilitate solving specific problems; MLG (Marketing Driven Growth Model) 1. Typical products: The more typical products include sharing economy products such as ofo and Mobike. In addition, I like to list retail products such as Yuanqi Forest and Coca-Cola. Although they are not considered Internet products, they are indeed a typical case of marketing-driven growth. Such products must also exist in Internet products, but there are relatively few; 2. The core driver of user growth: The driver of user growth of this type of product depends on brand marketing, and their growth relies heavily on the Internet or offline brand advertising, etc.; 3. Things to do at the core of driving growth: brand marketing, marketing advertising and promotion; 4. What users value more: users are more willing to consume quality, concept, spirit or culture, and pay more attention to spiritual value; If you are also thinking about how to achieve growth of your product, you might as well think about which one of the growth-driven models of your product are mentioned earlier. After determining this, the subsequent design of your growth strategy should focus on this. Key points: 4 common user growth-driven models on the Internet
3. Confirm the core leverage of user growth(1) What is the core leverage for user growth? The so-called core leverage refers to a growth strategy that has the highest cost-effectiveness, relatively small resource investment, but can bring huge returns, such as clearly clarifying the key indicators of user growth, user roles, growth sources, target users or usage scenarios. On the one hand, when exploring growth strategies, we should not let go of the near and far. We can pick some low-branch fruits first. Quickly seeing the effect of growth will also help us build confidence in growth; on the other hand, our resources are always limited. If we only focus on doing a few things, what should we focus on? This is the "core leverage". (2) How to determine the core leverage of user growth? In the case of Tencent meeting, we used the user growth formula model to clarify the core leverage of user growth through three steps. Next, we will explain in detail how to confirm the core leverage of user growth through the user growth formula model from a more general perspective. In order to facilitate practical operation, I have created a template to guide everyone to get a conclusion step by step: Step 1: Define the product’s user growth formula and confirm the key user roles of growth First of all, we need to define the user growth formula of our own product. At this step, we need to think about the following product users, which of the key user roles are, and then try to put the key user roles into the growth formula. In addition, try to use multiplication as much as possible for the growth formula, because the effect of multiplication is always stronger than addition. 1. What user roles does your product ecosystem include? Common product user roles include the following 5 types: user, initiator, manager, influencer, and purchase decision maker. Taking many products with social fission attributes for C-end as examples, the user role of their products usually includes two roles: initiator and user, such as social product products such as WeChat groups and QQ groups, collaborative products such as Tencent Conference and Tencent Documents, or group purchase products such as Pinduoduo. The core key figure of this type of product is the initiator. They drive user growth, and the initiator is the key user role. On the C-end products, usually the manager and purchase decision makers are the users themselves who use the product, and the influencers are more of the friends recommended by them. Some products do not have social fission attributes, so there may be no roles like the initiator, only the user themselves. For many B-end products, their roles will be undertaken by different users, such as the company's office software, whose users and initiators are company employees, but their managers are company administrators, while purchasing decision makers may be the company's heads, and influencers may be channel agents. At this step, you need to take a look at which user roles are involved in your product based on the specific situation of your product. 2. Try to put user roles into the user growth formula and confirm the best user growth formula For example, if a product whose role includes initiator and participant, its user growth formula can usually be: MAU=Initiator Size x Average number of users that an initiator can bring In addition to the examples of Tencent meetings before, let’s take Pinduoduo and Tencent Documents as examples:
In addition, some products that rely on channel business systems have user growth formulas that can be summarized as: MAU=Number of channel providers x Average number of users that a channel can bring Of course, if the user role of your product is very simple, you can also try to determine the calculation method of user growth formula through other methods, such as by user terminal (MAU=IOS terminal MAU+Android terminal MAU+PC terminal MAU), channel source (MAU=channel A source MAU+channel B source MAU+····), your goal is to find a user growth formula that best represents the characteristics of your product and business. Step 2: Drill down the growth formula and confirm the direction of user growth strategy In this step, you need to further disassemble your growth formula. Common disassembly methods include drilling down according to the dimensions of user activity, target population, and usage scenarios; 1. Disassemble according to the dimension of user's life active state, think about which one is the focus of growth, retention, and return? Usually we can divide the user into 3 parts according to the life cycle state:
Users of any product can be divided according to the above three types. What you need to think about next is that the core of your product growth at this stage depends on adding new users, retention or return to users; of course, if you have enough energy and resources, there is no problem in doing three together at the same time, but we always emphasize that at the same time, resources should be focused as much as possible in the most important place, and not cover everything. My opinions on this issue are as follows: 1. First look at the space scale: if you don’t have many old users, it doesn’t make much sense to retention and return. The focus will definitely be on attracting new products and solve the problem of survival first; 2. Secondly, when the space scale is comparable, look at the value of the unit user: for example, if the value of retaining an old user is higher than the value of attracting a new user, then the retention or reflow should be improved first; 3. Finally, when the value is equivalent, depending on the difficulty, it is usually easier to retain an old user than to acquire a new user. Therefore, for products that already have a certain scale of users, I recommend giving priority to retention; The specific situation can be analyzed based on your business and product development, but no matter what, you need to be clear: where to concentrate resources will determine the focus of the entire growth team; 2. Further drill down from multiple dimensions such as target population and usage scenarios; After you have confirmed which one is given priority to add, retain, and reflow, you can think about which target group or usage scenario should be entered from. In order to better analyze this point, I suggest that as we mentioned in the Tencent meeting case before, establish a four-quadrant analysis model, find out the parts that need key focus, and refer to the example of our scenario in Tencent meeting earlier. Step 3: Through the ICE scoring system, prioritize the growth strategy direction and determine user growth leverage What is the ICE scoring system? The ICE rating system is a strategy model proposed by Sean, the author of "Growth Hacker", to evaluate how to determine the highest priority when you have many growth strategies and ideas. This model mainly prioritizes diverse choices from three dimensions: influence, confidence, and simplicity. 1. Impact Assessment Influence means how much impact a certain strategy can have on overall growth. Usually, through the four quadrants of user value, influence is evaluated from two dimensions of growth space and growth value, and the core target population and usage scenarios are confirmed. 2. Assess confidence Some strategies may have a lot of room for growth, but it is not so easy to achieve. You need to evaluate your confidence in achieving your goals based on your own situation. 3. Simplicity Assess the simplicity of implementing a strategy; From the above three dimensions, we use the table template below to rate your growth strategy separately, then calculate a comprehensive score to determine the priority of the strategy, thereby confirming what the core leverage of user growth is. Key points: How to determine the core leverage of user growth? 1. Define the user growth formula for your product and put the key user roles into the growth formula; 2. Drill and disassemble the growth formula to confirm the priority of attracting new products, retention, and return, as well as key target groups and usage scenarios; 3. Through the ICE scoring system, determine the core leverage of user growth based on influence, assessment confidence and simplicity; 4. Find magical numbers for executable user growth(1) What are magic numbers? As we mentioned earlier, even if we have found the core leverage of growth, it is just one direction. For those who implement execution, it is still a "slogan". We need an indicator that can be implemented. The significance is that as long as this indicator is achieved through execution, the purpose of core leverage can be achieved. We call this Magic Number. This concept was not made up by me. This concept comes from the genre of growth hackers. It is widely used in major products in the Internet circles at home and abroad, including Facebook, Twitter, Tencent, etc., and are all using this concept. (2) Classic magical digital application cases Case of famous Internet companies For example, Facebook's product managers found that if a new user follows more than 7 friends within 10 days, he can stay and continue to interact. Under this guidance, all operation personnel or product personnel have a clear goal of promoting growth, which is to find a way to make a new user follow at least 7 friends within 10 days. If this is done, it will not be difficult to retain users. For example, in the Dropbox product, users will have a significant increase in loyalty as long as they use the folder function once, so the key to retaining users is to find ways to let users use the folder function once; in the same case, there are twitter and linkedin. There are many similar cases in domestic products, such as the magic number of Toutiao is to click more than 5 content, so its core growth goal is to find ways to increase the penetration rate of users with clicks more than 5 content. My work experience case The concept of magic numbers is a concept that I highly admire and has a very good guiding role. I used magic numbers in Tencent and later startups and found a more practical growth direction; for example, in a startup company after leaving Tencent (they provide online design tools and monetize through paid members), I found that users of this product may no longer be willing to pay in the future as long as they download more than 5 free templates for free, because they are used to free, so I told our team that we must let users pay before downloading 5 free templates. Case of startups providing consulting services For example, a friend from a jewelry retail store in mine, we discovered that if a user who enters the store has more than 3 products to watch after entering the store, the user will have a higher probability of paying; because I told her that you only need to tell them during the training of store employees to find a way to let users who enter the store watch at least 3 products in the display window. (3) How to find magic numbers? Regarding the process of finding magic numbers, I also try to create a template to help everyone: The process of finding magic numbers can be summarized in 3 steps: user research, logical assumptions/analysis of the importance of influencing factors, and correlation analysis of influencing factors. Before this, we first clarify which core leverage of user growth is based on to find magic numbers. Different leverage directions may correspond to different magic numbers. For example, in the previous case of Tencent meeting, we took the example of improving the retention rate of the teacher population. Step 1: User survey to determine WHY YES and WHY NO You need to answer the two questions of why yes and why no through user research. For example, in the case of Tencent meeting, we want to improve the retention rate of the teacher group, so we must first ask users what the reason for their willingness to stay and what the reason for their unwillingness to stay is, so that we can infer what the key factors affecting retention rate are. The key purpose of this step is to try to sort out the main factors affecting retention rate. Of course, if you have enough knowledge of the business, you can also sort out the influencing factors based on your own understanding. User research is just an auxiliary means. Step 2: Sorting logical assumptions and influencing factors' importance Based on the previous user survey, we can obtain some logical assumptions of factors that affect the target indicators. For example, the reason for affecting the retention rate of teachers' population may be the number of times they preside over the meeting. The logical assumption stage is a bold assumption based on the previous qualitative conclusions. In addition to qualitative assumptions of feelings, if your team has strong analytical skills, you can also analyze what factors are influencing and sort the importance of influencing factors through data modeling (including linear regression models, logistic regression models, decision tree models, shape value). Data experts who are good at data modeling and analysis can solve this problem more professionally. The principle is that when other factors are consistent, the impact of marginal changes in a single factor on the target indicators is evaluated according to the degree of influence; but this depends more on the team's data analysis ability. Many large companies have professional analysis teams to do it, but ordinary startups may not be able to do it, but this does not affect it. You may not necessarily need to be so strict, and judgment based on your own feelings may not work. Step 3: Correlation Analysis, Finding Magic Numbers Next, we need to verify whether our logical value is correct through data. This process is called correlation analysis. Correlation analysis is actually not complicated. We need to build a data relationship diagram of target indicators and influencing factors, see if there is a linear regression relationship between the two indicators, and then try to draw a relationship curve between two indicators. For example, we analyzed the relationship between the number of meetings hosted by teachers and retention rates. After building a relationship curve, we need to find the inflection point in the change curve from the curve. The so-called data inflection point means that when the influence factor exceeds a certain threshold, the marginal growth of the target indicator suddenly increases, and it manifests as a turning point on the relationship curve. Usually this turning point is what we call a magic number, which means that as long as the influencing factors reach this threshold, the target indicator can produce a significant increase; for example, in the previous case, we found that when the teacher group hosts N meetings within 30 days, the retention rate will be significantly improved, so we define the times as the magic number of retention rate growth. Key Point: How to Find Magic Numbers
5. Design user growth strategies based on magical numbersAfter confirming the magic number that achieves growth, what we need to think about next is naturally how to achieve the goals behind the magic number and how to achieve its growth strategy. Next, I will share common methods and models for designing growth strategies. In order to facilitate everyone to master and understand, I have made a template as follows: (1) Let users complete the golden experience path of the product once, and accelerate the experience of Aha moments ①What is the golden experience path for products? The golden experience path of the product is the main operation path when users use the core services provided by your product. Take Tencent Meeting as an example. The golden experience path of the host is: Home → Book a meeting → Share the conference invitation → Join the conference → Open the video/Open the microphone/Share screen. ②What is the Aha moment? Aha refers to moments when users feel very pleasant and have great satisfaction on the golden experience path. For example, when the host sees the successful booking page during the meeting booking, it will be very comfortable, or when the host successfully joins the meeting and the student responds and hears the teacher's voice. No matter your growth strategy, to achieve user retention and growth, you must find a way to let the user experience the product's golden experience path first, and find a way to make the user's Aha moment continue to magnify. Next, I will share 2 common accelerated product golden experience paths and Aha moment models: user journey map and emotional curve, growth hacker model, as follows: Model 1: User Journey Chart and Emotion Curve In this model, we first summarize the golden experience path of the user using the product in the user behavior module, and then describe the emotional performance of each link on this path in the emotional curve part, represent different emotions with different expressions, and then draw it into an emotional curve, and explain the reasons for each emotional expression in the emotional touch point part, so that the user journey chart and emotional curve are drawn. In this model, we need to focus on two key points:
The key to optimizing experience based on user journey maps and emotional experiences is how to improve the product experience at the trough points and how to further amplify the product experience at the peak experience points. I believe everyone can do these, so I won’t explain it too much. Model 2: Transformation = Desire - Friction "Conversion = Desire - Friction", this is a growth model mentioned in Sean's book "Growth Hacker". Many Tencent's growth experts will use this model to think about how to accelerate the golden experience path and Aha moment. The idea provided by this model is that if you want to accelerate the golden experience, you either find a way to increase users' desire to use the product or reduce friction on the user experience path; usually I recommend starting with reducing friction first, because this will be relatively simple. Common ways to reduce friction include: 1. Shorten the path: design the MVP path of your product to minimize unnecessary links; 2. Optimize the experience of key churn links: Through technical optimization or product optimization, improve the experience of a key churn link caused by high time and poor physical examinations, such as shortening the loading time and reducing the bug rate; Secondly, find ways to enhance users' desire to use products. Common ways to improve desire include: 1. Design stage rewards, allowing users to gain a sense of accomplishment in the stage: For example, when the registration process of many products is relatively long, they will prompt prompts such as "Congratulations on your success" on the registration path; 2. Simple design interaction, lower cost of understanding: the lower the cost of understanding, the stronger the desire for use, so it is very important to provide simple and easy-to-understand interactions; 2. Use addiction model to keep users using itWhat we discussed earlier is how to let users use your product first. Next, we want to think about how to make users continue to use it. The thinking model we often use here is the HOOK model, also called the addiction model. The idea that this model has improved is that if you want users to continue to use your product, you can start from these 4 angles: 1. Find effective trigger points and enhance triggers: Use effective triggers to reach users accurately and frequently, so that users can remember you and have the urge to use them. The triggers here include the contact push in the site, new media reach channels, online and offline advertising, etc.; one thing to remember is that you should not overuse triggers to make users bored. First, it needs to be accurate, and second, it is necessary to control the frequency, because it is usually difficult to achieve the effect of one touch. We need to trigger multiple times in a regular way to make users want to use, but the frequency needs to be controlled. 2. Guide users to take action: Don’t forget to remind users to take action. Many people’s operational experience shows that there will be obvious differences in the effect of having prompt actions and not having prompt actions, just like the click rates of buttons and no buttons in the advertising banner are significantly different; 3. Design rewards for users: Reward the user's use and efforts, so that the user can get a sense of satisfaction. The rewards here include physical rewards and spiritual rewards, such as recognition or interactions of others. 4. Design asset precipitation and cost investment: In product design, find ways to allow users to deposit some assets in your products, such as storing files, statistics, social relationships, etc., or find ways to make users pay some costs, including paid recharge, or users spend time filling in information, etc. Key point: How to design a user growth strategy? 1. Let users complete the golden experience path of the product once, and accelerate the experience of Aha moments
2. Use addictive models to keep users using: enhance triggers, guide actions, design rewards, and increase investment; 6. Verify growth strategy through AB experimentsBased on the above series of processes, we have obtained specific strategies to promote user growth. Next, what we need to do is to verify the effectiveness of the strategy. Regarding how to verify the effectiveness of the growth strategy through AB experiment, I will output a separate article to share this topic in the future. If you really need it, you can chat with me privately. To be simple, the main points that need to be focused on in AB experiments are as follows:
Finally, after we have verified that a certain strategy is a positive and effective strategy, we need to do two things next: 1. Improve the coverage of positive strategies and allow positive strategies to have a greater impact For example, in the previous experimental stage, the forward strategy only increased by 10% grayscale. After verifying that the forward strategy is obtained, you can increase the strategy to 90%. 2. Retain the experimental and control groups for continuous observation of effects Usually I would recommend that you continue to keep 5% of the experimental group and 5% of the control group to continue to observe the comparison of the effects of the strategy for a long time to monitor whether the strategy is always positive, because sometimes your positive strategy may not always be positive. Product changes and changes in the internal and external environment may affect your effect. Even in many products within Tencent, it often occurs that the current positive strategy will no longer work after a period of time. OK, at this point, the complete content of this article ends here, I hope it can help you. Author: Sanbai has something to say, official account: Sanbai has something to say |
<<: Four major marketing trends for private domain operations in 2024
>>: New "Spring Festival Gala Economics": No red envelopes, just want to sell goods?
Recommend
Private domain traffic solutions for different enterprises
Internet traffic has peaked, but private traffic c...
How to make more money from TikTok? Remember these 12 words
Is it difficult to make money on Douyin? How can y...
Is Douyin Shelf worth doing?
Douyin has invested 10 billion yuan to support e-c...
[Revealing] How do top bosses attract fans with their personal stories?
We often see some top bosses share some exciting a...
B station "horizontally"
Bilibili changes the playback volume indicator, an...
Is it difficult for a brand to break through the bottleneck? The boss’s cognition is the fundamental problem!
Brand growth requires founders to improve their co...
How to screen out high-value users from tens of millions or even millions of users?
When you have accumulated tens of millions or even...
The number of fans has skyrocketed 100 times, and "Dinosaur Carrying Wolf" also carries traffic
Recently, a brainwashing song called "Dinosau...
Can I change the name of an Amazon FBA shipment? Can I change the logistics information?
If Amazon sellers are self-delivering orders, they...
What is the Amazon Operations Tutorial? How to operate it?
As the Amazon platform continues to grow, more and...
Why shopping makes people anxious
As the end of the year approaches, there are more ...
AI live streaming chaos frequently occurs, and the platform takes action to rectify the chaos of "AI beauties" and selling courses
With the rapid development of AI technology, the l...
From Haidilao to Pangdonglai, has service become a new marketing trend?
This article tells the reasons why Pang Donglai be...
Where can I add SKU images on Amazon? What are the requirements for SKU images?
As one of the world's largest e-commerce platf...
Brand No. 1: Product power, brand power, and reach, three forces in one
This article analyzes the difficulties of brand ma...