How does product operation estimate product daily active users (DAU)?

During this time, the following questions are often asked:

1. Based on the current promotion and retention, what is the maximum daily active users we can achieve in the future?
2. In order to reach the planned daily active users in three months, we need to promote less every day?
3. How much retention does a product need to achieve in a certain country before promotion and revenue can be justified (ROI is positive)?

In fact, these questions are essentially answering one question, namely: How to estimate the daily activity of a product?

There should be many solutions to the problem. Here is a simple idea, summarized as follows.

Daily active users will be affected by many factors, including product iterations, operational activities, promotion changes, etc. Of course, some of these factors have less impact, while others cannot be predicted for the time being. Therefore, in the prediction process, we can eliminate some factors that have little influence, thereby simplifying it to a calculable state. (This process of simplifying to a computable level is actually called mathematical modeling.)

Therefore, in order to calculate, we first build a simple mathematical model of daily activity.

Build a mathematical model of daily activity

Among the factors that affect daily active users, the two most essential ones are the number of new users added daily and the retention rate of new users.

The daily active users on a certain day can be regarded as the new users on that day, plus the new users retained on the next day of the previous day, plus the new users retained on the second day of the day before yesterday...

By analogy, we can consider daily active users to be "the sum of new users on that day and the retained users on that day from the new users on each previous day" . Based on this, we can use a very simple formula to express daily active users.

DAU(n)=A(n)+A(n-1)R(1)+A(n-2)R(2)+… …+A(1)R(n-1)

Among them, DAU(n) is the daily active users on the nth day, A(n) is the new users on the nth day, and R(n-1) is the retention rate of new users after the n-1th day. If we assume that the number of new users per day is a fixed value A, the formula can be simplified as:

DAU(n)=A(1+R(1)+R(2)+… …+R(n-1))

The above formula can be seen as a simple mathematical model of daily activity. From this model, we can see that the newly added A is a relatively certain value, and the other part:

1+R(1)+R(2)+… …+R(n-1)

Determining the retained sum is a little tricky. You can use the following method to estimate retention.

How to estimate retention

Retention rate is the most core indicator of a product. The following figure is a retention rate attenuation curve of a product.

1-30 days retention rate decay curve

From the figure, we can see that the retention rate attenuation curve is very similar to the power function curve. In fact, the retention attenuation curves of most products in the industry basically conform to the power function curve.

Based on this, we can use the power function to approximate the attenuation curve of the retention rate, and thus smoothly estimate the sum of retention required in the daily activity model.

Generally, before estimating the retention of a product, we will have some prior data basis. If your product has been online for a while, you can use historical data as a basis. If the product has not yet been launched and there is no historical data, since the retention and decay rates of different types of products are different, the approximate retention data of similar products in the industry can be used as a reference for fitting predictions.

Therefore, retention curve fitting basically encounters two situations:

1. Now that we know the retention rate for a few days, how can we estimate the subsequent retention rate?
2. We don’t know the specific daily retention figures, we only know the retention data such as daily retention, weekly retention, and monthly retention, and we can estimate the retention figures for each day.

These two situations are essentially the same problem. Here we take the second situation as an example and briefly explain how to operate it. There are many methods for curve fitting. Here I introduce the simplest way, which is to use Excel to do a simple fitting calculation. The specific steps are as follows:

step1

Suppose we know the next-day retention, 7-day retention, and 30-day retention of a product as follows:

Retention of a product for a certain number of days

step2

In Excel, write the retention rate according to the corresponding retention days and draw a scatter plot:

Retention scatter plot

step3

Add a trend line to the above scatter points in the Excel chart, and in the trend line options, select the power function and choose to display the power function formula

Fitting curve based on scatter points

The resulting power function is:

y=0.4861*x^(-0.435), where x is the corresponding number of days and y is the retention rate of the corresponding number of days.

step4

Based on the obtained power function formula, the retention rate of all corresponding days can be calculated.

Calculate the estimated daily activity

Based on the power function obtained, after calculating the corresponding retention rate, you can simply sum it up and bring it into the daily activity formula:

DAU(n)=A(1+R(1)+R(2)+… …+R(n-1))

In this way, we can use the estimated daily new users to get the level of daily active users on the nth day in the future.

Summary & Postscript

By simplifying the daily active user model above and only considering the impact of new additions and retention on daily active users (other influences also indirectly affect daily active users through new additions and retention), we can roughly estimate the daily active user scale of the product in the future.

Then, based on the scale of daily active users, we can estimate some potential revenue, operating costs and other data. The above calculations are bound to have errors and cannot satisfy all scenarios, but the overall idea can be used as a reference and should be able to solve most related problems.

We made a small program that inputs data such as new additions and retention to directly estimate daily activity. The calculation method is the same as that in this article.

Related reading:

1. Product operation and promotion: How to compete for traffic?

2. How can product operations increase the number of new users and retain them?

3. Product operation: 2 major ways to get started to accurately capture private domain traffic!

4. Product Operation | How do stranger social products guide users?

5. How can product operations conduct good competitor research and analysis?

6. Product operation and promotion | 5 underlying ideas for traffic growth!

7. Product operation: application of data system under the growth model!

8. How can product operations conduct good competitor research and analysis?

Author: Nancun Xiaofu

Source: Nancun Xiaofu