How to turn new users into old users?

As online education enters the second half, refined operations centered on retention are the key to improving the competitiveness of educational institutions and education operators. I hope this will be of some inspiration to you.

When talking about product growth, many people focus on traffic, but in fact, more attention should be paid to retention rate, that is, the proportion of retained users to new users at that time, because only with a high retention rate can a product achieve long-term and stable growth, especially the growth of user scale.

Qu Hui, the author of "Silicon Valley Growth Hacker's Practical Notes", once gave such an example in the book: Suppose Company A and Company B start from scratch, Company A has a monthly retention rate of 80%, and adds 5 million new users per month; Company B has a monthly retention rate of 95%, and adds 2.5 million new users per month. After 6 months, Company A is still ahead of Company B, and after 3 years, Company B will surpass Company A.

This is the compound interest effect of retention, which will ensure that the party that is at a relative disadvantage in development has the opportunity to overtake and exchange time for space. The prerequisite for achieving this strategic goal is to have a higher retention rate. If Company A and Company B are replaced by any other company, the same principle applies.

Therefore, retention rate is one of the core indicators that need to be paid attention to for product growth. So how can we ensure the retention rate? Just grasp one thing - turn new users into old users.

Old users are not necessarily paying users, but rather long-term active and sticky users, who use the product more frequently and for a longer period of time. Therefore, to turn new users into old users, we must start by increasing their usage frequency and extending their usage time.

If you want new users to be willing to use your product frequently and for a long time, you must let them experience the core value of the product and the "aha moment", which is the key driving force for converting new users into old users. In addition, there are several other driving methods, and the author summarizes the following three.

Personalized recommendations

The so-called personalized recommendation is to continuously provide content and products that each user likes based on their needs or interests. The premise of the operation is that the user portrait is clear enough and the user needs are clear enough.

When we want to understand a user in depth, there are two main ways to do this: one is to have in-depth one-on-one communication with the user, and the other is to perform algorithmic analysis based on big data of user behavior. The former method is mostly used for the long-term maintenance of C-end users and B-end users in the high-unit-price consumer class. Because there is a strong trust relationship between each other and the grasp of user needs is relatively accurate, there is a high probability of transaction when recommending new products. The latter method is the main retention method for content-based apps and transaction-based apps, such as the content algorithm recommendations of Toutiao and Douyin, and the shopping cart product algorithm recommendations of Taobao and JD.com, all of which attract customers to continue reading and placing orders through precise content and product push.

Next, the author takes content apps as an example, starting from users, and briefly analyzes the process of improving retention rate through personalized recommendations.

➢ Select the content tab

Generally, during the activation stage, the system will guide users to select content tags of interest. This is the premise for personalized recommendations and the starting point for data recording. For example, if I register an account for a content app and select the three tags of history, comedy, and technology, when I enter the official interface, I will see several recommended contents based on these three tags, most of which are high-quality. In this case, eye-catching titles, keyword-containing titles, and attractive pictures are all key factors in attracting the author to click on the content.

➢ Record browsing behavior

As long as the user clicks and browses the article, the system will start recording the user's behavior, such as recording the speed at which the user reads the article, the frequency with which the user clicks on related articles, and whether the user likes, shares, comments, or collects these articles.

➢ Optimize recommendation model

After recording user behavior, the system will convert it into data and input it into the algorithm model. At the same time, because user behavior is constantly being recorded, data in more dimensions will continuously optimize the algorithm model, and more content that meets user needs will emerge, allowing users to spend more and more time using the product.

In fact, the rise of Toutiao and Douyin is due to the growth effect brought by personalized recommendations, especially Douyin, which has become one of the Internet's large-scale traffic pools in just a few years and has become a battleground.

When I first came into contact with TikTok, I was instantly attracted by the videos with cool special effects recommended by the system. Later, because I often browsed such videos, the system continued to recommend similar videos, causing me to spend several hours browsing every time I opened TikTok. I believe many friends have had similar feelings. It can be seen that the algorithm's personalized recommendations can allow users to invest enough time in the product, which will inevitably promote the improvement of retention rate and provide possibilities for monetization.

Refined recall

Refined recall is one of the most basic and reliable means to improve retention rate. When doing user segmentation, we often rely on user segmentation models. The commonly used user segmentation models are as follows.

• Clustering model: Divide a certain layer of users according to a certain dimension, such as grouping registered users by grade or region. In addition, you can also group according to multiple dimensions, often based on two dimensions, draw four quadrants, and then define the user attributes of each quadrant.

• Pyramid model: Users are stratified according to indicators such as business processes or participation. After stratification, the overall distribution of users is pyramid-shaped. For example, according to the "download → registration → payment → repeat purchase" model, users can be divided into new users, interested users, paying users, loyal users, etc.

• RFM model: Take the time of the most recent user behavior (Recency), the frequency of user behavior (Frequency), and the “total benefit” (Monetary) brought by the user behavior, divide the user levels based on these three dimensions, and classify and set operational indicators accordingly.

• Life cycle model: The user life cycle is divided into five stages: novice stage, growth stage, maturity stage, decline stage, and churn stage. According to the characteristics of each stage, operational goals and strategies can be designed for users at different stages.

The above are several common user stratification models suitable for improving retention rate. Next, we will focus on how to use user stratification models for refined recall and improve retention rate.

Refined recall is to perform detailed operations on user stratification and processes to ensure the overall recall effect. Its logic is very simple, namely setting goals → dividing users → finding problems → determining strategies → iterating processes.

Next, we will use activity incentives as a recall strategy and take the improvement of the retention rate of a reading app as an example to explain how to conduct refined operations based on this logic.

• Set goals: Guide existing users of reading products to participate in activities according to the process and achieve recall, with the goal of increasing the DAU of reading products.

• Divide users: Select the user stratification model listed above for stratification, such as analyzing users through the life cycle model, and obtaining 5 valid user levels based on the length of time the product is used, and then labeling and managing users based on the data.

• Find problems: Observe the number and labels of users at different levels, and analyze the actual needs and characteristics of users at each level. For example, novice users are not familiar with reading products and have low stickiness.

• Set strategies: Set up targeted activities for different levels of users, such as using normal promotional strategies for growth-stage users, using a “small discount + new book listing notification” discount strategy for mature-stage users, and using a “big promotion + high-frequency push recall” strategy for churned users.

• Iterative process: Design a specific recall process based on the strategy, and use data tracking to verify the effectiveness of activity incentives, especially in terms of nodes, copywriting, layout, paths, etc., and make timely adjustments and optimizations based on data results.

Regardless of the user stratification model, this logic can be used to design specific and refined recall strategies to improve the product's retention rate.

Design task system

In addition to personalized recommendations and refined recall, there is also a retention strategy that can improve user stickiness, which is to design a task system.

Its principle is actually very simple, which is to break down all related steps into multiple small tasks. Every time a user completes a small task, he or she can get accumulative virtual rewards, which encourages users to continue completing tasks until a habit is formed, thereby increasing the frequency of user use of the product and extending the time users invest in the product.

Why can the task system help users form habits? It is mainly based on a classic product operation model: trigger → action → variable rewards → investment.

➢ Trigger

The so-called trigger is to get users to use your product, but triggering requires inducement, which is the main driving force to attract users to use the product. There are many kinds of triggers, both visual and auditory, both external and internal.

Triggers caused by external factors are called external triggers, and triggers caused by internal factors are called internal triggers. Many times, the user's usage behavior is "activated" by external triggers. For example, you saw an eye-catching poster in your circle of friends, and the text on the poster read "Learn PPT layout design in 1 hour". At this time, you are worried about the design problem of PPT, so under external triggers, you scan the code to view the specific content and "start" the usage behavior.

Under the task system, sending recall information through SMS, App push messages, public account template messages, etc. is the most common external trigger method. When users see the task information sent by these channels, they often click out of curiosity, thus entering the stage of performing key behaviors.

➢ Action

Action is the second step in the model. The so-called action is the behavior and movement made by the user out of certain expectations, and this movement often occurs under the guidance of key behaviors. Taking the posters in the circle of friends as an example, scanning the code to view the specific content is the action taken by the user after being triggered.

As for the design of the task system, each specific task that users complete is the action link of the model, such as signing in, sharing, reading, answering, etc. Among them, any behavior specified in the task must have simple instructions, guidance or jumps, because users are willing to take action mainly for two reasons: one is that the behavior is simple, easy to operate, and does not require a high learning cost; the other is that the user has the subjective willingness to take action. Once these two motivations are violated, the task system will lose its effect.

➢ Variable rewards

As mentioned earlier, users have expectations when they take action, which means we have to respond to this expectation in the corresponding link, that is, to give certain feedback to user behavior, that is, rewards.

First, rewards are variable. This feature is reflected in the process of users taking actions, which makes users feel that there are constant surprises. For example, when you scan the code to view the course content, you find that the content described in the outline is very consistent with your needs, while you expect it to be similar to most outlines you have seen before. This is a variable reward.

For the task system, users will feel a variety of rewards when they complete tasks, such as:

• Reading an article that happens to be of high quality is the first kind of reward;

• After completing the task, the system will give corresponding virtual rewards such as 50 points, which is the second kind of reward;

• When checking the points rewards, you will find that the accumulated points can be used to redeem prizes or participate in draws, which is the third type of reward.

After the combined stimulation of these three rewards, users are more willing to proceed to the next task until all key behaviors are completed. Therefore, when designing a task system, the form of reward is crucial, as it determines the user's next action and subsequent use of the product.

➢ Investment

Investment is the last link of the model, which means letting users pay something. Only by letting users pay something for the product can the next trigger, action and reward be initiated, allowing the model to complete the cycle.

Typical forms of investment include: free courses, paid orders, virtual recharge, prize redemption and article collection, etc.

When designing a task system, connecting the points system with mall redemption, cash recharge and other processes, as well as giving out product-related rewards and rights after completing the task, are all effective means of getting users to invest.

By using the framework of this model for analysis, we can discover the underlying logic of how the task system helps products improve retention rates. As long as we grasp the key points from "trigger" to "investment", we can effectively improve retention rates.

The above is the underlying logic and related strategies for turning new users into old users in the retention phase. When we operate different types of products, we can refer to the relevant content and design more effective strategies.

Author: Wild Operation Community

Source: Wild Operation Community (dugu9bubai)