Cohort Analysis

What is Cohort Analysis?

• Just another tool
• A metrics analysis technique
• Key to avoiding vanity metrics and devloping actionable metrics
• Concept: Look of metrics of separate groups of users or customers
• Surgically measure the impact of changes, pivots: experiments

Picking the right stages is KEY

• Look at the lifetime a customer and identify key stages
• Make sure the stages can be measured and are associated with a specific user
• Make sure that they are meaninful to the business and the engine of growth
• Measure a baseline with current product
• Changes can now be measured to see if they had desired impact
• Identify key stages of relatioship with customer, e.g. AARRR

Step-by-step:

• Decide what “experiment” you want to look at
• Decide what “actions” you want to focus on
• Cohort: all customers who signed up during a certain week (time period)
• Each customer starts and stays forever in one cohort
• Create a generalization of the traditional funnel report
• Note that Cohorts can be non-time based too (e.g. employed vs. unemployed)

• Note that Cohort analysis is usually for the experiment, over a particular time period

• Cohort Analysis Example
• Engine of Growth Example