Cohort Retention

Purpose

Analyze retention with explicit cohort logic so the result actually means something.

When to use

Use this skill when:

• the question is about return behavior over time
• a team is using retention language loosely
• you need to compare cohorts, channels, or product changes

When not to use

Do not use this skill when:

• the task is about a one-time conversion flow
• the return rule is not meaningful for the product or business model

Required thinking discipline

• Define the cohort entry rule.
• Define the retained behavior.
• Define the time interval and retention horizon.
• Distinguish retention from reactivation and expansion.
• Evidence constraint: Every conclusion must cite specific data — a number, a rate, a segment, or a timeframe. Do not speculate without evidential basis. If data is insufficient, state what is missing rather than asserting an unsupported inference.

Workflow

1. Define the entity: user, account, buyer, team, or merchant.
2. Define cohort membership.
3. Define retained behavior and return window.
4. Choose the horizon and cadence.
5. Compare cohorts across useful segments.
6. Interpret differences carefully with lifecycle context.

Output format

• Cohort definition
• Retention definition
• Horizon and cadence
• Cohort comparison summary
• Segment differences
• Caveats and likely drivers

Good example

Cohort = first-paid-month account cohorts. Retention = account records at least one paid invoice in each subsequent month. Compare 6-month retention by acquisition channel.

Bad example

Retention is how many users keep using the product.

Why this is bad:

• cohort entry is missing
• return rule is missing
• time cadence is missing

Practical notes

• Say whether you are showing classic retention, rolling retention, or returning activity.
• Be careful when product usage is irregular by design.

Optional variants

• Marketplace businesses: distinguish buyer and seller retention.
• B2B SaaS: decide whether retention is user-level or account-level.