LTV calculation method and application

User life cycle is a concept that every product needs to pay attention to. Do you know how to calculate and apply it? This article will share this with you. Let's take a look.

This issue shares with you " LTV, customer lifetime value" . The mystery lies in the fact that everyone seems to understand its concept, but no one seems to be able to calculate it accurately.

Simply put, it refers to the total commercial value that each user, whether a buyer, user, or member, may contribute to your product or service in the future. It is worth noting that it measures the value of a single user and is used to indicate the quality of this user, not a scale indicator.

For example:

Assume that the customer acquisition cost of a reading product in the app store is 10 yuan , and it can generate an average of 0.5 yuan in revenue per day through various means such as user paid reading, ad clicks, membership purchases, etc. We have shared in previous content that the new users of a product remain at around 30-40% , which means that 60-70% of users will be lost. From the perspective of commercial monetization, there is a significant gap between customer acquisition cost and monetization revenue, which seems ridiculous.

- but

If you think so too.

You are totally wrong.

Capital is very smart and they will never do business that will lose money.

Back to the above question, now we extend the time period to 90 days, 180 days, 365 days, or even longer. Taking 90 days as an example, during these 90 days, the total value LTV that each user can create for the company is 0.5 yuan/day multiplied by 90 days, which is 45 yuan. In this process, the longer the number of days the user is active, the higher the profit. So it seems that spending 10 yuan to buy a user is still very cost-effective.

Therefore, LTV refers to measuring the value created by a user over a longer period of time.

That said so much.

What role does LTV play in actual business?

In actual application, LTV can evaluate the long-term value of users. By deeply analyzing key indicators such as user consumption, purchase, monetization behavior, and user retention rate, the expected value of users throughout their life cycle can be calculated to understand the long-term contribution of each user to the business.

Estimate the cost recovery period and the future DAU trend of the product;

Determine the cost of acquiring customers, compare the quality of users through channels, and adjust the delivery strategy. It is difficult to ensure the quality of users while pursuing user scale. You can achieve a reasonable ratio of high-quality and low-quality users based on the quality of channel users.

Calculate the ROI ratio and verify the profit model;

Evaluate the effects of product function updates and iterations, event operations, etc.

The following figure is a common derivation formula. The left side uses the daily retention rate multiplied by the daily average user revenue, and the right side uses the total life cycle length multiplied by the average revenue per user.

So which one is correct and how is it derived?

In fact, these two formulas are 1 formula, and the formula on the right is a simplified version of the formula on the left.

According to the definition of LTV , LTV = the total value of the user's lifetime = the total revenue of the user during the cycle divided by the total number of users.

The total number of users does not need to be estimated, so the question becomes: how to calculate the total user revenue.

Total user revenue can be calculated by multiplying the number of daily active users by the average daily active user revenue. Substituting this into the formula

The number of daily active users divided by the total number of users equals the daily retention rate.

When ARPU is equal to a constant, according to the definition, we can deduce again that LTV=LT×ARPU .

Students who want to study in depth can learn about derivatives and definite integrals in calculus. To simply extend this a little, the derivative describes the change of a function near a certain point, while the definite integral describes the cumulative effect of a function over a certain interval. They respectively describe the rate of change of a function at a certain point and the cumulative effect of a function over a certain interval.

Many of you may be confused after seeing this.

are you ok?

I was not good at math when I was in school.

Derivatives and definite integrals. I'm not good at math. Can I still learn this?

Is there a simpler way to calculate it?

The answer is: Yes.

OK, next, let’s learn how to calculate LTV using EXCEL.

Before calculating LTV , you need to process the " LT " and retention data.

LT = the average life cycle of a user. Note that it is different from LTV . LT lacks the word "value" compared to LTV . It is the number of active days from the first visit to the last visit of a user.

Let’s talk about retention data sampling. When calculating LTV, many students will directly take the retention data of the most recent 1-30 days, or directly calculate the average. However, when calculating the average, the weight should not be ignored, and the average should be converted into a weighted average.

If the scores are directly averaged without considering the weights, the average will appear too large, while the weighted average can better reflect the overall level.

New users are a variable. Suppose that the number of new users suddenly drops one day, but the retention rate happens to be at a relatively high level, the mean retention rate will be pulled up. However, after weighted calculation, the data from the abnormal date will not have much impact.

This is the error caused by data in two different calculation methods.

Data samples are generally retained for 1 day, 2 days, 3 days, 4 days, 5 days, 6 days, 7 days, 14 days, and 30 days. In theory, the more complete the retained data sample, the lower the data error.

Note: The data samples used above are all random numbers and have nothing to do with actual business.

Glossary:

LV "Life Time, the average life cycle of users"

LTV (Life Time Value)

Based on the retention rate calculated by weighted sum, a scatter plot is drawn, and a trend line is set using the scatter plot to obtain the calculation formula.

When setting a trend line, you can use exponential, linear, logarithmic, power functions, etc. to fit the curve. Generally, power functions are used more frequently. After confirming the type of function, get the formula through EXCEL, substitute the estimated number of days to calculate the retention rate, then check the display formula to display R² . These two fields will be used next.

The principle of LT to find the definite integral of the retention function is to obtain the shaded area under the curve, and the shaded area is the user life cycle LT .

The closer R² is to 1, the better the curve fitting effect is.

Assuming that the ARPU value is stable over a period of time, you can directly use the average value and use the ARPU value as a constant.

Assuming that the ARPU value is unstable, it is necessary to set a piecewise function based on the ARPU value within the data acquisition period.

After calculating LT, when using LT×ARPU to calculate LTV , you need to fit different retention curves based on the estimated target. Different channels and payment types will result in very different user retention rates and ARPU values. It is possible to retain one variable in the analyzed data to reduce interference items.

For example.

The formula is like this: distance = speed * time .

The advantage is that it is easy to understand.

The disadvantage is also very obvious, that is, it is too theoretical.

In real life, vehicle speed changes at any time due to factors such as power, traffic, and weather.

Similarly, LT is not an easily estimated data, and ARPU is not an easy-to-average constant value. Such calculation results can only be used as a macro reference.

Substituting retention rate and ARPU value into the LTV formula yields:

The calculation of LTV is often discussed together with " CAC " and " PBP " . Taking the above figure as an example, assuming that the user activation cost CPA = 15 yuan, the cost can be recovered in 180 days for new users from 5/1 to 5/30. After 365 days of registration, each user can earn an average profit of 23.5 yuan .

Glossary:

ARPU (Average Revenue Per User) is calculated as follows: ARPU = Revenue / Number of Active Users, i.e. total revenue in a certain period / number of users in that period. This user can be a paying user, an active user, etc.

PBP (Payback Period) generally refers to the time period it takes to recover the investment costs.

CPA (Cost Per Action) generally refers to download/activation;

CAC (Customer Acquisition Cost) generally refers to paying users. Many people do not understand the difference between CPA and CAC and often confuse them, or call them both customer acquisition costs. In fact, they are completely different indicators. Traffic does not necessarily become users. There is a paid conversion rate CR in between.

This issue mainly shares the derivation of common life cycle formulas and how to estimate user life cycle using EXCEL tables.

It should be noted that LTV is a static fitting data. In actual business, the estimated data will have certain errors. It is necessary to combine the actual situation of the product and make continuous corrections to reduce the data error.