Use this "retention rate model" to calculate how far your product is from making a million a day?

Almost all operational work is carried out around the four links of " attracting new customers ", " retention ", "activation" and " conversion ".

We often face this problem: we attract a group of users with great difficulty, but they all leave as they play, not to mention conversion and monetization.

Therefore, it is very important to understand the " retention rate ". It is an important indicator to measure product quality and operational effectiveness, and it is also an important reference for estimating product profitability!

This article has 3,200 words.

1. What is “retention rate”?

The retention rate refers to active users. Specifically, it refers to users who have completed a certain behavior within a certain period of time. For example, a content platform can set its own standards by referring to the number of opens, usage time, interaction, and content output. It should be noted that a single user may complete a specific behavior multiple times. When counting, it is necessary to deduplicate according to the user dimension, that is, only count as one user.

For example, a group of new users come in one day. Some of them use it for a while and never come back, some continue to use it a few times and then leave, and some can continue to use it actively for a certain period of time. We call those who remain active retained users. The ratio of retained users to new users in this batch is the retention rate.

When considering the retention time, there are generally the next day, the 7th day, the 30th day, etc., which correspond to the next-day retention rate, the 7th-day retention rate, and the 30th-day retention rate respectively. There are also retention rate calculation forms based on a whole week or month, such as next month retention rate.

(7-day retention and weekly retention are easy to confuse. Let's take an example to make it clear: if 100 new users are added on January 1, and 20 of them are still active on January 8, then the 7-day retention is 20%, that is, the 7-day retention refers to the number of new users on the first day who are active seven days later; if 1,000 new users are added from January 1 to January 7, and 300 of them are still active between January 8 and January 14, then the weekly retention is 30%, that is, the number of new users in the first week who are active in the second week.)

Based on the existing number of retained users , the retention rate can be calculated, and based on the retention rate, the future number of active users of the product can be predicted.

As an operator, don’t simply assume that all users you bring in are your users, and don’t always use the “total number of users” as an indicator to be complacent about. Instead, you should pay more attention to the number of active users and retention rate.

Because only truly active users can generate commercial value.

2. Retention Rate Model

To understand the retention rate, this article will use a magic weapon - "product profit model based on retention rate".

Model 1.0

In version 1.0, retention rate is just a preliminary model: (actually it is just an addition formula)

(Retention rate model formula)

This model explains that the number of active users comes from "new users within the stage" plus "retained users in each previous stage." For example, if you want to calculate the retained users for 7 days, then the total number of active users on the 7th day = the number of new users on the 7th day + the number of new users on the 6th day x the retention rate on the 1st day + the number of new users on the 5th day x the retention rate on the 2nd day + … + the number of new users on the 1st day x the retention rate on the 6th day. If we calculate it on a weekly or monthly basis, the same principle applies, so to put it simply, this model is the result of adding up the number of retained people in each stage.

Let's explain it with an example:

For example, if the number of active users of a product on July 1 is 500, and then the daily active numbers of these 500 people in the next 7 days are pulled, the 7-day retention rate of the product can be calculated:

(Retention rate on the day = number of remaining active users in the batch / number of active users on July 1)

In this way, we get the stage retention rate of the product. Then we can calculate the daily retention of natural new users based on this retention rate. If the number of new users on July 1 is 1,400, we can continue the calculation:

(Number of remaining active users on that day = retention rate on that day * number of new users added on July 1)

Therefore, we have preliminarily obtained the changes in the number of active users on July 1st in the following days. Of course, the reality is that there are new users every day, so we can calculate the retention rate of new users every day, and then add up the historical retention to get the total daily retention.

If there are 500 new users on July 1, and the number of new users increases steadily by 500 per day, then based on the retention rate, the total number of active users on July 7 can be calculated:

Model 2.0

Once we can calculate the number of active users in the future, we can roughly calculate future revenue, sales and other data.

This model can be further optimized by adding more parameters to make the calculation results closer to the actual situation.

It should be noted that there are many factors that affect the model results. The more factors are considered, the closer the model is to the actual situation and the higher the accuracy. But at the same time, factors will affect each other, which increases the complexity of the model, increases the time required for calculation, and makes the calculation results more difficult to obtain. When using the model, please think carefully about the balance between these two points.

So the question is, what factors need to be considered in reality? What could be added to this model?

1. Users come from different channels , so different channels need different retention rates;

2. As operational activities progress, monthly natural additions will also change;

3. Many products will do paid promotion every month, which needs to take into account the conversion rate of the promotion ;

4. The retention rate of users attracted by promotions and activities may be different from the retention rate of naturally added users ;

5. Considering the user life cycle , historical stock users will have a fixed churn rate;

6. Among the total active users, a certain percentage of them pay;

7. As the product value increases, the payment conversion rate of new users will increase every month;

8. With the refined operation of the product, the consumption capacity of a single user will gradually increase during the life cycle;

9. You can also estimate the total revenue based on the per capita contribution rate;

…

Well, if we add all these parameters, the model is already quite complicated. But as long as the formula is correct, the result will always be close to the actual situation.

Even though you may not understand it at first, we still YY a product, YY some data, and made cost calculations and revenue estimates for it.

It can be seen that the iOS channel and the Android channel are calculated separately, and several parameters are added, including the historical user loss (10%) and the monthly natural increase of 2%. More importantly, the conversion rate and cost of paid promotion are added.

The ultimate move is coming!

Now that we have the users and costs of the two channels, we can calculate more data! First calculate the total cost (blue part), then calculate the user data (green part), and then, you can add your own conversion data (yellow part) for measurement! Even GMV can be calculated!

Pay attention to the red background of the last line. If you add KPIs, you can directly know whether you need to adjust your strategy!

Of course, all the data here are made up by me. If you really think that we can achieve nearly 20 million GMV in November, then you are really cute.

—— Pay attention——

Ahem, the homework is here!

After understanding this model, here is an application problem. Please hand it in to the class representative before the end of get out of class. Do not whisper to each other!

A certain product invested 4.5 million from its launch to the end of June and gained 2 million users, including 1.2 million IOS users and 800,000 Android users. The old IOS users were lost at a rate of 10% per month (this month's old users are 90% of last month's old users), and the old Android users were lost at a rate of 15% per month (this month's old users are 85% of last month's old users).

In July, the product added a monetization function. It is known that the payment rate of active IOS users is 20%, and each paying user contributes 2 yuan to the company per month; the payment rate of active Android users is 15%, and each paying user contributes 1.5 yuan to the company per month. It is known that the number of new IOS users in July was 500,000, and the number of new Android users was 400,000. If the current number of new users per month is maintained and without considering its own costs, in which month can the company start to make a profit? (Attached is the historical retention rate, assuming that iOS has 1.2 million active users and Android has 800,000 active users as of June)

The question looks long, but it is actually not difficult! (My math level is only up to addition and subtraction within 100, and I need a calculator for everything else...)

As a reader, do you know how to count?

The results are posted here. There are two solutions!

Solution 1

（iOS user estimate table）

This table calculates the monthly active users and monthly revenue statistics of the iOS terminal, and adds several new indicators such as historical stock user churn (10%), payment rate (20%), and average contribution (2 yuan);

(Android user estimate table)

This is the estimated situation on the Android side, and the calculation method is the same as above. In the last two rows we add up the total revenue from the two channels and subtract the costs, and it is clear that the product will start to be profitable in November.

Solution 2

I won’t analyze the specific steps of Solution 2, but you should be able to understand it from this table.

If you understand it, forward it to your boss, and then you can brag about it.

If you don't understand, change your career as soon as possible~

3. What is the use of this model?

There are three main aspects:

1. For start-up products, you can estimate the time it takes for the product to become profitable and adjust the strategy in a timely manner

For start-up products, whether it is the boss, product manager or operator, the most feared thing is not only the unstable product functions, but also the unstable users and unclear benefits. With this table, it can be easily calculated.

2. Split the KPI of each parameter

For products that have passed the startup phase, you can intuitively see future results under the current operating status based on the target KPI. If the target is not achieved, timely adjustments must be made. More importantly, you can set KPIs for each channel, payment rate, average order value, and total cost, which will be much easier after the splitting.

3. Show off

It’s up to you to decide, just don’t go too far, or you’ll get slapped in the face by your boss.

Of course, you have to understand that this model only has a measurement function and its accuracy is limited, so the calculation results are only for reference, and the calculation formula only helps operators enhance their understanding of users.

When it comes to improving retention, there is no one-size-fits-all method that can immediately increase product retention. What we should focus on is whether the product itself meets user needs well. The retention rate is not high, and 95% of the reasons are that the needs of the product itself are not well met.