Customer Cohort Analysis for E-commerce: Initial steps toward personalization

To read the full article, please visit Datarize Blog : https://blog.datarize.ai/en/customercohortanalysis

Do you think Gen X, Millennials, and Gen Z experienced puberty in exactly the same way at the age of 17? Given the distinct differences between each generation, it's reasonable to assume that their experiences were unique. Even after 10, 20, or even 30 years have passed, these experiences contribute to shaping them into distinct adults.

Even if they are at the same age of 17, the 'type of experience' each one has will differ. Therefore, rather than grouping everyone simply as '17-year-olds,' we should instead examine characteristics by dividing them into groups based on 'generational 17-year-olds' who have had 'similar' experiences. Customer cohort analysis operates on this principle, tracking how a group with 'similar experiences' evolves over time.

Why Customer Cohort Analysis is essential for Ecommerce

Although new members who signed up in July and October performed the same action of signing up, each group likely had a different experience. Factors such as differing marketing strategies or changes in the service's UI/UX between these months could account for this. Most crucially, the products offered might have varied as well. For instance, members who joined in July might have been inclined to purchase short-sleeved shirts, whereas those who signed up in October could have opted for cardigans. Therefore, while the act of signing up matters, the timing of their registration can also have a substantial impact on various indicators.

Two things that can be learned from customer cohort analysis

1. With cohort analysis, you can assess the performance of the action taken.

For instance, imagine one group receives a $5 off coupon upon signing up, while another is offered an event where they can purchase a popular item for just 10 cents. You can then compare the groups based on the specific incentive they were given over time. If a group's performance is consistently strong, this indicates effective marketing. However, if there's a decline in performance after the initial phase, it suggests an increase in the proportion of cherry pickers.

?

2. With cohort analysis, you can identify churn patterns and take steps to prevent churn.

Additionally, you can predict the churn timing of new members by identifying their group type. By addressing their needs at the predicted point of churn, proactive measures can be taken to prevent them from leaving.

?