Why are you losing users? There are some ways to improve user retention!

In the Internet industry, we usually attract new customers, but after a period of time, some of them may gradually leave. Those who stay or frequently visit our company website/App are called retention.

Nowadays, people often use "daily active users" (DAU) and "weekly active users" (WAU) to monitor websites, and regard the gradual increase of "daily active users" over a period of time as a good phenomenon. If retention analysis is not done, this may be a wrong start:

For example, a company has carried out many new customer acquisition activities, bringing in many new users, and the number of active users is also increasing. However, the excessive number of new users also masks the problem of high churn rate. In fact, customer retention is gradually decreasing. If the high churn rate cannot be prevented, the new users will always leave one day.

The users who remain are the real users. If you want to achieve sustained and real growth, you must find ways to keep users.

Taking SaaS companies as an example, the cost of acquiring users is huge both in time and money, and may take two to three months. Take the left picture below as an example. We spent more than $6,000 to acquire a customer. The customer paid monthly according to the agreement, for example, $500 per month:

The upfront costs are high, and we may only recoup them after customers use the product for a year or two. If this customer has been lost before, it means that our product has lost money and has not even recovered the original investment.

The longer users use the product, the higher the cash flow or profit it will bring. Not only SaaS companies, but for any Internet company, the longer users stay, the more valuable they are. This is the core meaning of retention.

From the above picture we can see two points:

First, the duration of use. The longer the product stays in our store, the better.

Second, we hope the profit is as high as possible. How can the profit be as high as possible? The higher the retention rate , the larger the profit area.

If we do a good job of retention, users will continue to use our products, continue to generate value and bring profits.

Having introduced the concept and significance of retention, let’s move on to today’s topic: how to improve our retention curve through retention analysis. Take the following figure as an example:

If the retention rate of our product is the green line at the bottom of the above picture, the vertical axis is the retention ratio and the horizontal axis is time. After one day, only 35% of the 100% new users we attracted stayed. On the 7th day, the number became 20%, and then slowly decreased. After the 60th day, it reached an effect of about 10%.

Can we improve it gradually by making improvements in certain aspects?

If we let the green retention line rise to the orange line and then to the red line, the first-day retention rate will be as high as 70%, and the seven-day retention rate will be more than 60%. By 60 days and 90 days, the retention rate will be as high as 60%. This means that of the 100% of the people we attracted through market acquisition, 60% of them stayed after 90 days. At the beginning, looking at the green line, our 90-day retention rate is 10%. If we can reach 60% through our efforts, it will bring us a steady stream of wealth and cash flow income.

The same is true for e-commerce and content social industries. Facebook can achieve $60 billion in advertising revenue per quarter because of its huge user base.

Today, I will give you some ideas through some retention analysis methods, and see how to improve our retention rate by optimizing our products. Below is a common retention curve. I divide it into three parts: the first part is the oscillation period, the second part is the selection period, and the third part is the stable period.

During the oscillation period, the number of new users decreases dramatically in the first few days, and after a few days it drops to a few percent or even lower. This is called the oscillation period.

After the oscillation period comes the selection period. Generally speaking, users have a preliminary understanding of the product during this period and begin to explore our company's products to see if the product meets some of their core needs. If they are satisfied, the customers are likely to stay; if they are not satisfied, the users will leave.

Next comes the stabilization period, when the retention rate enters a relatively stable stage.

Different strategies should be adopted for different periods. Generally speaking, during the oscillation period and the selection period, attention should be paid to the retention of new users. After entering the stable period, focus should be placed on the retention of product functions.

Next, I will talk about retention analysis using two real cases. Before that, let’s first talk about the steps of retention analysis. Of course, the basic steps of data analysis are similar:

Sidekick is a SaaS company that provides enhanced email functionality. Users can use its personalized templates to send emails to others, and can also monitor whether the recipient has opened the email.

The three retention curves in the figure below represent the user retention performance in the first week, second week and third week of December 2014. This is called data monitoring. The continued decline in the retention curve is a problem discovered through data monitoring.

Specifically, there are two problems:

1. The retention rate in the first week dropped significantly;

2. In the second week, the curve we expected did not appear (the retention curve tended to stabilize). We hoped that customers would stay, but it was still decreasing. A decline means that one day all the new members we attract will leave.

At this time, two goals were set:

1. We hope that the retention rate in the first week will increase and will not continue to decrease. If you are a 2C business, the retention rate on the next day may increase.

2. We hope that our retention curve will no longer decline and enter a flat pattern.

The target retention curve should look like this:

Through continuous practice and analysis, the company eventually maintained its retention rate above 20%. So how do they do it?

Goal 1: Improve first-week retention

They have discovered the problem of decreased retention rate, but they don’t know the cause of the problem. Next, they need to explore the data to find the reason for the decrease in retention:

We first segment users into groups, and then compare the behaviors of retained and churned users. This company specializing in email found that among those who churned in the first week, 60% sent emails once on the first day, and only about 20% sent emails twice. In other words, 60% of users used the product once and then left.

So next we have to ask why these users leave? So I made a pie chart based on user feedback, and it was clear that there were two major problems:

The first 30% of people do not feel the value: this means that the product does not generate value and they want to uninstall it.

The second 30% of people do not understand the purpose of the product: I downloaded this product, but it is not what I thought it was.

These two groups of users account for 60%. We often say that we should listen to the voices of users, and the first thing we should solve is the problems of these 60% of people. Therefore, we can assume that users have not discovered the value of our products quickly.

1. Cut off the functions that are used less frequently

So the first thing they did was, since users couldn’t quickly discover the value of our product, they would cut off some of the complex and difficult-to-understand features first, and then see if the retention rate would improve. They found that the retention rate continued to decline and there was no improvement.

2. Prompt users to discover the value of the product

The second attempt was, since users didn’t know what the core value of our product was, I gave them hints. As a result, the retention rate continued to decline. This method didn’t work either.

3. Guide users through videos

The third attempt was, since users didn’t know how to use our product, we made a video. In fact, many companies were doing this, but the data showed that it still didn’t work.

4. After adding a reminder

They did more than 20 experiments before they found a feasible solution:

After users downloaded and installed the product, they wrote a sentence: You can go to your mailbox to use your email.

Their products have to be downloaded from the web page, but users use them on the client. Users may not think too much about it and think that since they have downloaded a plug-in on the website, they can just use it directly on the website without going back to the user's client. So they gave a tip: just go to your Outlook and use it. After adding this sentence, the retention effect changed a lot.

This is the result in May 2015. It was blue before and was finally upgraded to yellow.

Goal 2: Improve the retention curve

After the new user retention rate is improved, let's look at the second question. Our goal is to move the retention curve upward, which is in a stable period. For this problem, it is very important to split the product by function and check the retention of each function. We call it product retention analysis.

We need to understand the retention rates of all the different features in this product, which retention trends are decreasing and which are increasing, and analyze the features with decreasing retention rates to find out why.

For products, we use the methodology introduced above to optimize products; for operations , we perform operations to win back customers.

Through continuous trial and error, analysis, and monitoring, they improved and optimized the product features that reduced retention rates, thereby maintaining retention at a high level.

Whether we are focusing on retention or attracting new customers, it is basically the same. At the beginning, we monitor every core function we focus on. If we find any abnormalities, we start analyzing the problem. We can conduct various data explorations to find the problem; set a goal, such as hoping to increase the retention curve by 10 points within half a year (10 points is a very significant improvement).

Sidekick spent half a year from December to May, during which time continuous hypothesis, verification, analysis and observation led to a significant improvement in retention.

Once the retention curve is raised, we can monetize users and even do other things. We continue to attract new users, so that the user base continues to grow, and the retained users accumulate steadily. These are our most important users and can be monetized. As for those unstable users, we still need to make various product modifications, operations or marketing operations to gradually turn them into retained users and achieve real growth in the company's active users.