Growth Hacker's Guide: How to Improve User Retention

If you still invest huge budgets to attract new customers when user retention rate is low, then you are actually just renting traffic . In this case, how do we improve user retention ?

In the previous article "Step-by-Step Growth Hacker Guide: How to Improve User Registration Conversion and User Activation", we have shared how to improve user registration conversion rate from the funnel of " target users - copywriting content- channel selection-landing conversion". So after the user completes the registration conversion, the job of the growth hacker is to try to retain the user - the higher the user retention rate , the longer the user life cycle , and the higher the user life cycle value.

I mentioned in the article " AARRR is a thing of the past, and R AR RA is a better growth hacker model" that the AARRR model that focuses on user growth has lost its practical significance. When McClure proposed the "Pirate Metrics-AARRR Model" in 2007, the customer acquisition costs (CAC) were still very low, so the AARRR model emphasized that "customer acquisition/user acquisition (Acquisition) indicators" are the primary indicators.

But today, the market situation is completely different. If you look at the traffic prices of major advertising/ social channels now, the customer acquisition cost (CAC) is already ridiculously high, and the market situation is completely different from that in 2007. Therefore, if at this stage of development the top priority is still to emphasize customer acquisition, I think it is inappropriate.

So we need a better growth hacking model, and that model is the RARRA model.

RARRA Model

In the RARRA model, user retention is the most concerned factor - because user retention rate can truly reflect the value of the product. As I have always emphasized, if you still invest a huge budget to attract new users/acquire customers when your user retention rate is low, then you are actually just renting traffic. This is not real customer acquisition at all, because no matter how many new users you attract, they will eventually leave.

So how can we improve user retention rate?

Let me pour a bucket of cold water on you first. Improving user retention is definitely not as simple as teaching you a few martial arts moves. It is analyzed and driven by data. Therefore, before we take a series of operational actions to try to improve user retention, we need to conduct user retention analysis and group analysis first. We need to find out the user retention rate, at what point users churn, and why users churn . This is the basis for all your operational interventions. Then we can conduct optimization experiments in a targeted manner, otherwise it will be useless.

It is crucial to clarify these three issues:

1. What is the user retention rate for N days?
2. When did users churn?
3. What retention nodes cause users to churn?

To analyze user retention rate, we can try Cohort Analysis.

What is Cohort Analysis?

Cohort Analysis means that you group users according to their source or behavior to understand their retention on your product. Cohort Analysis is generally divided into two categories:

• Acquisition Cohorts: refers to the division of users when they first register for the product. They can be divided according to the acquisition date or the source of the acquisition channel .
• Behavioral Cohorts: Users are divided according to the behavioral trajectories they take in your product. These behaviors can be any behavioral events such as application launch, application uninstallation, product transaction, etc.

Let’s take a look at an example of how to use customer acquisition group analysis:

Example 1: Customer acquisition groups - divided by customer acquisition channel - user retention rate

Example 1 shows the acquisition cohorts divided by different acquisition channels. You can analyze the weekly retention, that is, the 7-day retention, from the table:

• Organic Search
• Directly enter the URL to access (Direct);
• Referral traffic (Refral): Tieba , friendly links, etc.
• Social Media
• Search engine bidding traffic (Paid Search);
• Email.

What is the role of Acquisition Cohorts?

You can compare different user sources and then filter out the best channels.

For example, you can find out from the above table that the users coming from organic search engines are the most (probably because SEO optimization is the best), but its weekly retention rate is very low. However, although the number of users who visit by directly entering the URL (Direct) is not as large as the natural traffic from search engines (Organic Search), the weekly retention rate is the highest (perhaps they have already developed brand trust in the product).

Although there are not many users coming via email, the weekly retention rate is relatively high, and we can increase marketing investment in this channel in the future. On the contrary, users who come through search engine bidding traffic (Paid Search) are not only small in number, but also have the lowest retention rate.

Therefore, when you adjust your marketing strategy later, you can reduce the search engine keyword bidding budget and invest the funds in email marketing , or polish the product to increase user stickiness and retention of users who directly enter the URL to visit (Direct).

Let’s look at another example:

Example 2: Customer acquisition groups - divided by customer acquisition date - user retention rate

Example 2 is the Acquisition Cohorts divided by acquisition date. You can analyze it from the table:

• On January 25 (Day 0), there were 1,098 new users. The retention rate on Day 1 was 33.9%, the retention rate on Day 2 was 23.5%, the retention rate on Day 3 was 18.7%, the retention rate on Day 7 was 14.5%, and the retention rate on Day 10 was 12.1%. That is to say, the retention rate of the 1,098 new users on January 25 was 12.1% on the 10th day, leaving only 132 users, with a user churn rate of 87.9%.
• On January 26 (Day 0), there were 1,358 new users. The retention rate on Day 1 was 31.1%, and the retention rate on Day 2 was 18.6%… The retention rate on Day 9 was 11.3%, and the user retention rate was as high as 88.7%. Only 153 of the 1,358 users remained.
• You can use this table to continue analyzing daily user retention rates for different acquisition dates…

What is the role of Acquisition Cohorts?

This chart can clearly tell us that the user churn rate on Day 1 is the highest. From January 25 to February 3, there were a total of 13,487 new users (Day 0), but the average retention rate on Day 1 was only 27%. That is to say, only 3,641 of the 13,487 users were left on Day 1, and the remaining 9,846 users all left, with a user churn rate of 73%.

So there you have it — users churn the most on the first day, and you don’t know why. Because customer acquisition group analysis will only tell you on which day the user churned, but it will not tell you the specific reason for the churn and the specific churn node.

Therefore, we need to use user behavior group analysis to gain insight into the reasons for user churn and the specific nodes of churn.

Let's look at an example:

Example 3: User Behavior Groups - Onboarding - User Retention Rate

From the above picture, we selected Onboarding as the observation item to see the user retention after completing Onboarding. A closer look reveals that the highest churn rate occurs on Day 1, and the average retention rate is between 23% and 27%. This means that after Onboarding, the churn rate is as high as over 73%!

This means that there is a problem with Onboarding. We need to improve Onboarding user guidance and tell users our product value, product features, etc.

What is the role of Behavioral Cohorts?

It can tell us why users churn and the specific nodes at which users churn. Behavioral Cohorts and Acquisition Cohorts complement each other.

If Behavioral Cohorts is not easy to understand, then the funnel analysis model can more intuitively show the user churn situation and make the user churn situation more concrete. Let's take an example:

Example 4: Funnel Model - Onboarding - Churn Rate

From Example 4, how many new users successfully completed all onboarding flows?

Let's analyze the complete process of the entire funnel: Download and install - Log in to the App - Guide page 1 - Guide page 2 - Guide page 3 - Complete user onboarding

Our analysis of this funnel model shows that most users give up when entering onboarding page 3 from onboarding page 2, and only 30% of users complete the entire onboarding process.

Is this funnel analysis model useful?

This funnel analysis model tells us that the reason for the serious loss of users is that there is a problem with the Onboarding page, and the problem is on Onboarding page 3.

So how should we improve it?

Onboarding Pages are very important for products. First, they describe the main functions of the product. Second, they convey the value of the product to users. Third, they help users use the product better. Therefore, if the Onboarding Pages are not attractive, the user’s favorability will decrease. It is recommended to conduct A/B testing on copywriting or visual design, and then test, analyze, and adjust again.

The above are just examples to illustrate that user behavior group analysis and funnel model analysis need to be used in combination, which can help you analyze the time of user churn and the user churn nodes more clearly and intuitively.

So through the above three points:

1. Acquisition Cohorts;
2. Behavioral Cohorts;
3. User funnel model analysisFunnel Analysis.

We can know:

• What is the user retention rate for N days?
• When did users churn?
• What retention nodes cause users to churn?

So what we need to do next is to carry out targeted operations on specific issues in order to truly improve user retention.

Data-driven growth is truly the key.

Source: Xi Wenyi