How to reduce product user churn rate?

Reviewing this project, I gained a lot. The most important thing is to summarize a set of operating methods based on data and rapid trial and error as action plans.

QuestMobile released the "Mobile Internet Panoramic Ecosystem Traffic Insight Report" in March this year. In this report, we can see that the growth rate of China's mobile Internet user scale had dropped below 5% in June 2018 , and in February 2019, the number of users increased by only 7 million compared with December last year.

The previous high-speed growth is gone forever, and the competition for traffic is becoming increasingly severe. With the disappearance of the demographic dividend, traffic is becoming more and more expensive. It is becoming increasingly expensive for companies to achieve growth by buying traffic through a large number of channels and adding new users.

When there is not enough cake on the plate, everyone can only grab from each other's bowls. The fierce competition for traffic can be clearly felt from the rising customer acquisition costs.

According to my understanding,

The average download and activation cost of a tool product is 1 to 5 yuan per person.

The average download and activation cost of a live broadcast product is 5 to 10 yuan per person.

The average download and activation cost of an online education product is between 10 and 50 yuan per user.

The average cost of downloading and activating an e-commerce product is between 50 and 200 yuan per user.

Compared to a few years ago, these prices are already astronomical figures (the customer acquisition costs and effects of each channel vary. The above data are only the data I know and do not represent universality).

Assuming that 2,000 new customers are acquired every day, the investment in promotion alone will range from thousands to hundreds of thousands of yuan. If you can't control your retention rate, the amount of money you lose every day will be terrifying.

Therefore, no matter what type of company it is, if it only continues to spend money on promotion and buying traffic, but does not pay attention to retaining users or caring about lost users, it is tantamount to chronic suicide. After all, money always runs out.

So you will find that in the past two years, there is more and more content about refined operations, and the term "refinement" is mentioned more and more frequently. Because refined operations actually mean paying attention to every bit of traffic and reducing losses at all costs through various operational means. The various operating methods and approaches derived from this, such as user stratification, user labeling, RMF model, etc., are all for refined user operations.

Not only on the user side, in terms of refinement, we actually have a lot of things to do.

For example, optimizing the user activation process, or using a strong guidance mode or a weak guidance mode based on product characteristics.

The example I’m going to give you today is an optimization project that reduces churn rate by optimizing user usage processes.

Case Review

The background is this.

The company has launched a new live broadcast product and is in a period of rapid development. The channel spent a lot of money on promotion and traffic, but data shows that user retention is very low.

Many users exit the product directly after registering and logging in, and never log in again, which means that the money spent on promotion is wasted, which gives leaders a headache.

So we temporarily formed a project team to solve this problem and improve retention.

1. Request data.

We organized the data of each step after new users entered the product in a certain period of time, made a funnel data page, and conducted an analysis. As shown below (because the data involves the company's business, all data in this article have been anonymized and do not represent the actual situation).

According to the data, we found that many users entered the product but the live broadcast ended before they even entered the room to watch it. Such users will definitely leave and will never come back, because they have not even experienced the Aha moment of the product.

1. Thinking and discussion

Based on the results of data analysis, we are thinking and discussing why these users left without even entering the room?

1. Is it because the traffic channels are not good and the new users are not accurate enough?
2. Is it because the brand is not well-known enough and users don’t know what kind of live streaming product our product is?
3. Is it because the process is too complicated, resulting in too much loss?
4. Or is it because the anchor selection interface is too complicated, causing users to be confused and leave?

In fact, there are definitely several reasons.

Traffic generation channels definitely require continuous trial and error, and then find the traffic source that best suits your product; building brand awareness also requires the efforts of the marketing department/brand department to gradually establish the brand awareness of the product; and if the process is too complicated, product operations are needed to optimize it.

These are all things we can and must deal with, but what should be prioritized and what should be delayed? How do you prioritize everything?

After internal discussions, we decided to prioritize optimizing the process first . Because this part of the content is easiest for us to control overall, and the easiest to quantify and compare data.

3. Once the discussion is over, start working immediately.

We first use a flow chart to represent the original usage process of new users.

Download - Activate - Register - Login - Enter the room

Here, we will divide the data into three groups for analysis and comparison.

1. Control group: No change to the process, still download-activate-register-login-choose to enter a room
2. Test Group 1: Change the user process to this: download - activate - register - log in - directly enter the popular anchor room
3. Test Group 2: Change the user flow to: Download - Activate - Enter the room directly - Register (if you want to speak or send gifts) - Log in (if you want to speak or send gifts)

4. After analyzing and comparing the three sets of data, we obtained the following feedback.

Test Group 1, by directly guiding new users into the room after logging in , allowing them to experience the product’s Aha moment at the first time, significantly improved the next-day retention of new users, by almost 15%.

Obviously, because users can enter the room directly after logging in, one step of user operation is reduced, which effectively reduces churn and improves retention.

Test Group 2, by directly guiding activated users into the room and omitting the registration and login steps , the retention rate of new users increased more significantly, by 30%.

There are two reasons for this growth. On the one hand, it reduces user steps, reduces churn and improves retention. On the other hand, the number of registered users has decreased, which indirectly improves retention.

Theoretically, the solution of test group 2 is more effective. But in fact, the leader finally decided on the plan for Test Group 1.

The reason is that although the total number of users is not a data that the project team needs to improve, it is indeed a very important data for a product. Therefore, the approach of increasing retention by reducing registrations is not feasible.

This also exposes a problem, which is that we need to look at the problem from a more comprehensive perspective. Only by taking a holistic view and looking at the problem from a higher perspective can we come up with better and more appropriate solutions.

At this point, this project has come to a temporary end.

Summarize

Reviewing this project, I gained a lot. The most important thing is to summarize a set of operating methods based on data and rapid trial and error as action plans.

The first step is to pull all relevant data for the problem encountered, analyze the existing data, identify various possible causes, and then make reasonable guesses.

The second step is to propose effective growth plan hypotheses based on our conjectures and assumptions. Clarify the priorities of various matters, establish process tables, and identify the most important variables.

The third step is to design A/B tests for the most important variables and conduct experiments.

The fourth step is to analyze the results of the AB experiment, verify our hypothesis, and then iterate the product. Then, based on the data of the iterative product, reverse engineer the iterative effect.

When it comes to data analysis, many people have a misunderstanding. The approach should be to obtain all data, conduct a comprehensive analysis, and identify problems through the data. Instead of looking for support in the data when you have a problem.

This method of data analysis is not comprehensive enough. After all, data is cold and heartless, and the patterns behind them are what we should observe and understand. Data analysis combined with business understanding is the right approach.

Whether it is traditional live streaming or e-commerce live streaming, it seems to be content operation, but in essence it is human operation. Without users, content is just cold and replicable; only people are unique and full of individuality.

Related reading:

1. How to reduce the churn rate of APP users?

2. User churn prediction model, how to evaluate its effectiveness?

3. How to improve product stickiness and reduce user churn rate?

4.10 strategies for APP to reduce user churn rate!

5. How to solve the high user churn rate? Here are 10 strategies

6How to analyze retention data and reduce user churn?

Author: T Brother

Source: T Brother