User Operations: The Value of User Retention to Growth

Nowadays, the cost of acquiring customers is getting higher and higher, and user retention is not so easy; but it is precisely because of the high cost of acquiring customers that retention becomes more important; the author of this article explains in detail the value of user retention to growth, and also expresses under what circumstances retention is suitable. Let’s take a look.

I believe everyone has done this problem when they were young: a reservoir has an inlet pipe and an outlet pipe; given the water flow rate at the inlet and outlet, calculate the amount of water in the reservoir after a period of time.

This is very similar to evaluating the effect of user growth. The water inlet is compared to attracting new customers, the water outlet is compared to customer loss, and the water storage capacity after a period of time is the retention capacity of growth.

After entering the second half of the Internet, the total number of users has become difficult to increase, and the cost of attracting new users has also increased a lot; under this trend, the importance of retention has become increasingly prominent.

1. What is retention?

1. Definition of retention

In the Internet industry, users often learn about and use new products through advertising, recommendations and other communication channels, and some users will no longer use these products for various reasons.

We define users who use our products for the first time as new users, and users who continue to use these products after a period of time are called retained users.

Retention is calculated based on the initial behavior time of a certain user group. It describes whether the same group of people who performed a certain behavior performed the expected behavior after a period of time. The initial/return behavior here can be any event. If not specifically specified, it generally refers to "launching the product."

When measuring retention, focusing only on "launching the product" often overlooks a lot of important information, so most companies also make retention statistics for some key paths based on their own business characteristics; for example, e-commerce platforms will count repurchases based on repeated purchase behavior, and securities companies will count investment transaction retention based on repeated asset transactions.

To ensure the consistency of statistical benchmarks, we generally use retention rate to express the retention effect, that is, the proportion of the retention volume to the initial volume of the same group; taking a certain information platform as an example, using the opening of the APP as the retention mark, its next-day retention rate is calculated as follows:

Assume initial behavior = launch product, return behavior = launch product

The number of users who launched the product on a certain day = A, and the number of these users who still launched the product on the second day = B

>The number of users retained on the next day in this group = B

>The next day retention rate of this group of users = B/A*100%

2. Retention is the foundation of growth

If we think of the number of users as water in a reservoir, the market will slowly inject new users into the pool, and some old users in the pool will inevitably become indifferent to the product and leave.

Observing the entire growth process, we can roughly summarize the following formula:

Effective user growth = new user acquisition - user loss = new user acquisition * retention rate

In other words, the significance of improving user retention is to ensure the effectiveness of user growth by reducing user churn.

Suppose there are two startups, A and B, located in a blue ocean industry and started at the same time. These two companies are the only ones competing in the target market. They have the same daily new users, but Company A's products have a 100% month-to-month retention rate, while Company B's products have only a 0% month-to-month retention rate.

We observe the trend of the total number of users of these two companies in months, and draw the following graph:

The analysis shows that when the product has a small user base in the early stage, the churn phenomenon has little impact on the overall situation, and the effective growth rate at this stage is determined by the strength of attracting new users; when the user base reaches a certain scale in the mid-term, daily churn will greatly drag down the effectiveness of attracting new users on that day, resulting in weak growth.

There are limited potential users. Once the market capacity ceiling is reached, the number of new users will approach 0, and the number of users will enter a one-way downward trend.

If the users lost from Company B are also included in Company A's potential new users, Company A's user base will continue to grow after reaching the critical point of 1/2 market capacity until the market capacity is filled (i.e., Company A's* curve).

2. How to describe user retention

The total number of users is the summary of the number of new users and the number of lost users. Therefore, simply observing the trend of user volume is not enough to evaluate the product retention situation.

We can draw a retention curve by calculating the retention rate of the same user group at different times, and track horizontally whether they continue to use the product, what proportion of them churn, and when they churn; thus, we can understand their retention situation and optimize the product.

1. Find the key behaviors

Analysys defines user retention as follows: Retention refers to the behavior of users who have used apps, websites, and other applications and continue to use them after a period of time.

The user retention here is broad. When the definition of "usage" is not specified, we often record "launching the product" as usage. Users who continue to use the product after a period of time will be recorded as retained users.

When measuring retention, focusing only on “launching the product” often overlooks a lot of important information. So what other behaviors can we observe to measure product retention?

Products are the medium through which enterprises exchange value with users in order to achieve commercial value.

Based on this concept, healthy products will have their own core functions to meet their business purposes. The effective operation of these functions is the core value of the product. Behaviors that allow users to feel the core value of the product and repeat them over and over again can prove that the product can continue to generate value for users, forming a virtuous circle and thus retaining them - these behaviors are the key behaviors we need to build retention rate.

For example, Product A is a weather software that not only allows users to check the weather but also allows users to record their moods. Assuming that all users only record their moods but do not check the weather after opening the software, the behavior of these users will interfere with the iteration of the product. At the same time, it is meaningless to evaluate the quality of weather products based on the size of this group of users.

The key behaviors here often refer to the operating results of the core business of the product. Taking product A as an example, a more reasonable approach is to record "user checking the weather" as the key behavior, and at the same time observe the retention of "users repeatedly opening the software to check the weather within a certain period of time."

2. Find the observation period

For a travel product, users may only open it once every two weeks or so; reading news is much more frequent than traveling, and the interval between users’ use of news software may be one day.

If the retention rate of products is uniformly calculated based on a one-day cycle, it is obviously inappropriate for evaluating the retention health of tourism products.

The above picture shows the weekly usage frequency of some apps in six categories, including weather, over the past year.

It can be seen that the usage frequency of different categories of applications varies greatly. For example, the frequency of weather, news, and financial applications is much higher than that of travel applications.

This is determined by the specific scenario to which the product belongs, so many companies often expand their business areas in order to increase the frequency of product use and improve retention rates by penetrating more user demand scenarios.

In addition to looking at industry benchmark data, we can also observe from the product's internal data monitoring system; take a longer time period, capture users who "use key behaviors more than twice", and look at the time distribution between their two behaviors, which is the current usage cycle of the product.

The usage cycle represents the standard interval for a user to use a product. Whether a user triggers a key event in a standard interval can be used to evaluate whether the user is retained within this cycle. We can use one usage cycle as the minimum observation unit and evaluate the trend of retention rate by observing multiple consecutive usage cycles.

3. Retention curve

A relatively basic method of describing retention is the retention curve. It uses time as the horizontal axis to depict the changing trajectory of user retention over time.

As shown in the figure below, the general retention curve will first drop rapidly (new registered users quickly lose), and then there will be three categories depending on the product:

Flat curve: This curve remains flat after a rapid decline, neither rising nor falling; it represents that within this user group, the number of users has reached a balance: some new users see the value of the product and return continuously for a long time; of course, a higher curve after stabilization represents a product that is more popular with users; but generally speaking, a flat retention curve can make people feel at ease because it means that the number of product users will not shrink.

Decline curve: This curve continues to decline without reaching a flat level, and will eventually reach a very low value; it represents the continuous loss of users in this user base: very few or even no users find the value of the product; at this time, we need to change the product and find a way to make some users become loyal users first, so as to flatten the curve.

Smile curve: This curve flattens out after a rapid decline and then slowly rises; it represents the continuous improvement of the product: more new users see the value and improvement of the product and return in the long term.

3. How to evaluate retention effect

1. No need to grow, no need to retain

As mentioned above, retention is the foundation of growth; in other words, if the product does not require growth, there is no need to put costs into building retention.

For example, consider the convenience store in a scenic spot as a product, tourists in the scenic spot (mostly outsiders) as potential users, and the core business as commodity trading, and the convenience store makes money through buying and selling.

Customers in scenic spots are basically one-time customers, so no matter whether they stay or not, it is difficult for them to make secondary purchases and increase business profitability. The cost-effectiveness of such retention is extremely low. Under the premise of sufficient costs, if the canteen wants to obtain higher profits, it should put more costs on competition with peers, such as purchasing more suitable goods for the scene, doing promotional activities, etc. to lower the competitiveness of other canteens in the scenic spot and maximize profits.

The same principle applies if we switch our perspective back to Internet products.

First of all, we need to make it clear whether our product has a high enough demand for growth, and then we can think about whether we should do retention and how high the retention should be.

2. Is the retention curve reasonable enough?

1) Is retention getting better?

User retention rate represents the proportion of product recognition among the same group, and it will become more significant over time.

Observe the changes in retention rate of the same group of users over time. If the curve tends to be flat, it means that the product does meet the needs of some users and makes them willing to stay.

If the curve goes down in one direction, then we need to consider whether there has been a failure in attracting new users and they have deviated from the target user group; or whether the product itself has not yet achieved market fit.

Through healthy product iterations, improved accuracy in attracting new customers, and other means, the product's retention rate should tend to stabilize earlier and increase its approach value.

Observe the user retention rate curves of different batches on the same path. Ideally, the most recent curve should develop towards a better shape.

2) Use the user lifetime value to infer a reasonable retention curve

As long as users "stay", they have the potential to realize value. Generally speaking, the longer the period of stay, the more marketing opportunities there are. Similarly, the higher the total life cycle value (LTV) that can be converted.

Here is a diagram of the stages of the user life cycle. You can see that the stages from user introduction to user loss are progressive. Users are often unable to create value before activation and after loss, so the peak of value conversion is from activation to active stage.

We take the user's retention time and the average monetization value per unit time, and we can get a rough user cumulative conversion value curve (S curve); it describes the conversion value accumulated by the user in each retention stage, and α and β are the dividing points before user cycle activation and after loss, respectively.

On this basis, we draw an average user retention line:

• Let Fa(x) = retention rate at a certain moment, Fb(x) = cumulative customer value at a certain moment;
• T(x) = new users * Fa(x) * Fb(x) = total conversion value after a period of time of attracting new users;

What impact do different retention rates have on the value of user conversions?

We try to adjust the user retention rate to observe its impact on the total value conversion of this group of users; from observing the graph, we can find that, under the premise that the user value curve remains unchanged, the retention rate curve passes through α~origin, and the total conversion value is no different. However, after crossing the value curve of the active period, the total conversion value can increase a lot every time the slope is raised to a certain level.

The reason is that the two curves before activation and after loss are in the flat zone of the S curve, and the fluctuation value is close to 0, so the change of the retention curve has no obvious effect on the product result; during the active period, the user value curve increases unidirectionally over time, so every time the retention rate curve is raised by a certain unit, the result will be significantly improved.

In addition to controlling retention rate, we can also influence the final cumulative total value from another perspective.

Keeping the retention curve unchanged, let’s try adjusting the value curve.

It is not difficult to find that even if the retention curve finally falls before the activation point, as long as we optimize the product and advance the activation point, we can still achieve the same effect.

IV. Conclusion

1. Retention is the basis for controlling the effectiveness of growth, and there is no guarantee that the increase in retention rate is only short-term growth;
2. Retention rate is a key indicator for observing retention performance. In addition to looking at retention, we also need to define "key events" and "observation cycles" to build a retention dashboard;
3. Not all products need retention rate. If it is meaningless to business value, you don’t need to consider retention.
4. Raising the retention curve and pre-product activation point as much as possible can effectively improve the user value conversion effect.

Author:caNn

Source: PM Qingfeng