Cohort Retention & Decay Analysis
Understanding how long customers remain engaged and when they tend to drop off is crucial for any business looking to improve customer lifetime value and reduce churn. Retention analysis measures the proportion of users from a given cohort—such as those who made their first purchase in a particular month—who return or remain active in subsequent periods. By structuring users into cohorts based on their initial interaction (as you did in the previous chapter), you can track retention rates over time for each group. This enables you to see how quickly engagement decays and to identify patterns in customer loyalty, which is vital for SaaS, e-commerce, and subscription businesses aiming to optimize customer journeys and maximize recurring revenue.
1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465import pandas as pd import numpy as np # Fixed realistic transaction data data = { "user_id": [ # Cohort 2023-01 users 1, 2, 3, 4, 5, 1, 2, 3, 1, 3, # Cohort 2023-02 users 6, 7, 8, 9, 6, 7, 6, # Cohort 2023-03 users 10, 11, 12, 10, 12 ], "order_month": [ # Cohort 1 purchases "2023-01", "2023-01", "2023-01", "2023-01", "2023-01", "2023-02", "2023-02", "2023-02", "2023-03", "2023-03", # Cohort 2 purchases "2023-02", "2023-02", "2023-02", "2023-02", "2023-03", "2023-03", "2023-04", # Cohort 3 purchases "2023-03", "2023-03", "2023-03", # FIXED this line "2023-04", "2023-04" ] } df = pd.DataFrame(data) # Assign cohort_month as the first purchase month df["cohort_month"] = df.groupby("user_id")["order_month"].transform("min") # Convert to datetime df["order_month"] = pd.to_datetime(df["order_month"]) df["cohort_month"] = pd.to_datetime(df["cohort_month"]) # cohort index df["cohort_index"] = ( (df["order_month"].dt.year - df["cohort_month"].dt.year) * 12 + (df["order_month"].dt.month - df["cohort_month"].dt.month) + 1 ) # pivot table cohort_counts = ( df.groupby(["cohort_month", "cohort_index"])["user_id"] .nunique() .reset_index() .pivot(index="cohort_month", columns="cohort_index", values="user_id") ) # retention matrix cohort_sizes = cohort_counts.iloc[:, 0] retention = cohort_counts.divide(cohort_sizes, axis=0).round(3) print(retention)
1234567891011121314151617181920import seaborn as sns import matplotlib.pyplot as plt # Assume 'retention' DataFrame is already computed as above plt.figure(figsize=(8, 5)) sns.heatmap( retention, annot=True, fmt=".0%", cmap="Blues", cbar=True, linewidths=0.5, linecolor="gray" ) plt.title("Cohort Retention Rates by Month") plt.ylabel("Cohort Month (First Purchase)") plt.xlabel("Months Since First Purchase") plt.yticks(rotation=0) plt.show()
When you examine the retention heatmap, you will notice that retention rates typically start at 100% for each cohort's first month, since all users are present at the beginning. Over time, these rates usually decline as some customers become inactive or churn. The rate and pattern of this decay reveal important business insights:
- A SaaS company may see a gradual drop each month, indicating steady but manageable churn;
- An e-commerce business might display sharp drops after the first or second month, reflecting a reliance on one-time buyers.
High retention in later months signals strong product-market fit and customer loyalty, while steep early declines may highlight issues with onboarding, value delivery, or product satisfaction. Recognizing these patterns allows you to target improvements where they will have the most impact on long-term revenue and growth.
Дякуємо за ваш відгук!
Запитати АІ
Запитати АІ
Запитайте про що завгодно або спробуйте одне із запропонованих запитань, щоб почати наш чат
Awesome!
Completion rate improved to 5.56Cohort Retention & Decay Analysis
Свайпніть щоб показати меню
Understanding how long customers remain engaged and when they tend to drop off is crucial for any business looking to improve customer lifetime value and reduce churn. Retention analysis measures the proportion of users from a given cohort—such as those who made their first purchase in a particular month—who return or remain active in subsequent periods. By structuring users into cohorts based on their initial interaction (as you did in the previous chapter), you can track retention rates over time for each group. This enables you to see how quickly engagement decays and to identify patterns in customer loyalty, which is vital for SaaS, e-commerce, and subscription businesses aiming to optimize customer journeys and maximize recurring revenue.
1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465import pandas as pd import numpy as np # Fixed realistic transaction data data = { "user_id": [ # Cohort 2023-01 users 1, 2, 3, 4, 5, 1, 2, 3, 1, 3, # Cohort 2023-02 users 6, 7, 8, 9, 6, 7, 6, # Cohort 2023-03 users 10, 11, 12, 10, 12 ], "order_month": [ # Cohort 1 purchases "2023-01", "2023-01", "2023-01", "2023-01", "2023-01", "2023-02", "2023-02", "2023-02", "2023-03", "2023-03", # Cohort 2 purchases "2023-02", "2023-02", "2023-02", "2023-02", "2023-03", "2023-03", "2023-04", # Cohort 3 purchases "2023-03", "2023-03", "2023-03", # FIXED this line "2023-04", "2023-04" ] } df = pd.DataFrame(data) # Assign cohort_month as the first purchase month df["cohort_month"] = df.groupby("user_id")["order_month"].transform("min") # Convert to datetime df["order_month"] = pd.to_datetime(df["order_month"]) df["cohort_month"] = pd.to_datetime(df["cohort_month"]) # cohort index df["cohort_index"] = ( (df["order_month"].dt.year - df["cohort_month"].dt.year) * 12 + (df["order_month"].dt.month - df["cohort_month"].dt.month) + 1 ) # pivot table cohort_counts = ( df.groupby(["cohort_month", "cohort_index"])["user_id"] .nunique() .reset_index() .pivot(index="cohort_month", columns="cohort_index", values="user_id") ) # retention matrix cohort_sizes = cohort_counts.iloc[:, 0] retention = cohort_counts.divide(cohort_sizes, axis=0).round(3) print(retention)
1234567891011121314151617181920import seaborn as sns import matplotlib.pyplot as plt # Assume 'retention' DataFrame is already computed as above plt.figure(figsize=(8, 5)) sns.heatmap( retention, annot=True, fmt=".0%", cmap="Blues", cbar=True, linewidths=0.5, linecolor="gray" ) plt.title("Cohort Retention Rates by Month") plt.ylabel("Cohort Month (First Purchase)") plt.xlabel("Months Since First Purchase") plt.yticks(rotation=0) plt.show()
When you examine the retention heatmap, you will notice that retention rates typically start at 100% for each cohort's first month, since all users are present at the beginning. Over time, these rates usually decline as some customers become inactive or churn. The rate and pattern of this decay reveal important business insights:
- A SaaS company may see a gradual drop each month, indicating steady but manageable churn;
- An e-commerce business might display sharp drops after the first or second month, reflecting a reliance on one-time buyers.
High retention in later months signals strong product-market fit and customer loyalty, while steep early declines may highlight issues with onboarding, value delivery, or product satisfaction. Recognizing these patterns allows you to target improvements where they will have the most impact on long-term revenue and growth.
Дякуємо за ваш відгук!
