A hypothesis about the effectiveness of a certain element

[Joywtome231]

1. Mailing efficiency
In the first example, let's imagine that an online cosmetics company decided to test how effectively email newsletters work, or more precisely, to identify the sources that bring in the most loyal subscribers. Mailings are usually sent through:

brand social networks;
form on the official website or application;
lead magnet on the promo page.
The main goal is to understand where exactly the company's most loyal customers and target audience come from. Cohort analysis will actively help with this.

We see that the values ​​for September were as follows: 1,000 people signed up for the newsletter via social networks, 1,850 via the form, and 1,400 via the lead magnet. Given all this data, we can create three groups — cohorts. Let's enter the indicators into the table.


Here, one cohort is one row in our table. Percentages by month are how many people remained subscribed to the mailing list after a month, two or three. For example, let's make a cohort of "social network" users and look at it. It is clear that a month after subscribing, at the end of September, 53% of the initial number of people did not unsubscribe from the mailing list. Or 1000 * 53% = 530 people.

Statistics show that the most loyal and interested users in the product come from the source “website form”: as many as 73% stay with the brand and read the newsletter in the first month and 64% after three months.

For people who subscribed to the newsletter via social networks, the retention and interest rates are lower. And at the same time, users who subscribed for the lead magnet leave almost immediately. They are not interested in the newsletter and the company's products: by the end of the first month, more than half of the users unsubscribe.

Thus, in this situation, it is worth focusing more on attracting valuable regular customers through the form on the site. And including improving, reformatting the attraction of users through social networks. It is better to abandon the latter method.

2. Cohort distribution
The second example will clearly show how you can distribute users into cohorts depending on the feature.

For example, on September 10, 25-year-old Mikhail, a resident of Novosibirsk, found a website selling books and textbooks through the Yandex search engine and purchased an economics manual there. This set of data makes it possible to immediately distribute this user into a large number of cohorts:


3. A\B testing and cohort analysis
A\B testing can test process, or even an entire marketing campaign. A classic iceland phone number list example: loyal customers are divided into 2 groups, and one is shown a site with a green “Buy” button, and the other with a “Blue” one. The one with the best conversion can be left.

However, to verify the result, you can move on to cohort analysis and conduct additional research.


Let's sum it up
Cohort analysis is an indispensable method. It can provide a business with a large amount of necessary and important information. For any businessman, such data is of great value: based on it, you can improve the sales process, create a portrait of an ideal client or adjust the advertising strategy within the application.

However, this tool should be used with caution. If not all data is collected and analyzed, or the information for analysis is incorrect, this can lead to negative consequences for the company and financial losses. That is why it is so important to clearly understand the basic essence of cohort analysis, know how to use it and what stages of implementation to follow.