retention-analysis

Retention Analysis: Measuring What Keeps Users Coming Back

Quick Start

/retention-analysis

Then provide:

1. Retention data (D1, D7, D14, D30 rates -- or I'll query your analytics MCP)
2. Product usage frequency (daily, weekly, monthly -- how often should users return?)
3. Known churn reasons (if any -- from interviews, support tickets, surveys)

I'll analyze your retention curve shape, identify the biggest drop-off, compare retained vs churned user behavior, and recommend interventions.

Output: Saved to thoughts/shared/pm/analyses/retention-analysis-[date].md Time: ~15 min with data, ~25 min with cohort deep-dive

When to use: When diagnosing churn problems, measuring product-market fit, or optimizing for stickiness

Framework source: Aakash Gupta's retention frameworks and "Ultimate Guide to Activation"

Context Routing Logic (Internal - for Claude)

Automatic Context Checks: When this skill is invoked, immediately check:

| Source | Files/Folders | Search Terms | What to Extract | | ----------------- | ---------------------------------------------- | -------------------------------------------------------------------------- | ------------------------------------------------------------ | | Metrics/Analytics | thoughts/shared/pm/metrics/.md | D7, D30, retention, churn, cohort, "monthly active", DAU, WAU | Current retention curves, cohort performance, churn rates | | User Research | thoughts/shared/pm/.md | churn, "stopped using", "didn't come back", "why I left", "why I switched" | Churn interview quotes, reasons users stop using product | | Meeting Notes | thoughts/shared/product/meeting-notes/.md | churn, "cancelled", "downgrade", lost deal, customer feedback | CS feedback on churn, customer complaints, drop-off patterns | | PRDs | thoughts/shared/pm/prds/.md | retention, sticky, habit, engagement, notification, reminder | Features built to improve retention | | Business Info | thoughts/shared/pm/context/business-info-template.md | target user, use case, frequency, engagement, core activity | How often users should use product, what drives stickiness |

Context Priority:

1. Internal context FIRST (business info, retention metrics, churn research)
2. Analytics MCP SECOND (if connected - query retention cohorts, churn reasons)
3. Framework guidance LAST (generic retention tactics)

Cross-Skill Links:

• If activation issues found → Link to activation-analysis (fix activation first)
• If expansion opportunity identified → Link to expansion-strategy
• If feature opportunity identified → Link to prd-draft

Step 0: Understanding Your Current Retention Reality

Before diving into retention analysis, let me check what data already exists about your users...

Checking:

• thoughts/shared/pm/context/business-info-template.md for expected product usage patterns
• thoughts/shared/pm/metrics/ for existing retention metrics and cohort data
• thoughts/shared/pm/ for churn interviews and user feedback
• thoughts/shared/product/meeting-notes/ for CS/support feedback on why users churn
• thoughts/shared/pm/prds/ for features built to improve retention

[If analytics MCP connected]: "Let me also query [PostHog/PostHog] for your current retention curves, churn rates by cohort, and behavioral differences between retained vs churned users."

Based on what I find, I'll show you:

Internal Intelligence Summary

From Business Info:

• [Your product's intended usage frequency]
• [Target users and their engagement patterns]
• Example: "Product designed for daily use by individual contributors in engineering teams"

From Metrics/Analytics:

• [Current D7 and D30 retention rates]
• [Retention curve shape (flattening, declining, or smiling)]
• [Churn rate and trends]
• [Retention differences by cohort or segment]
• Example: "D7: 42%, D30: 28%, declining curve, but organic cohort outperforms paid by 15%"

From Churn Research:

• [Top reasons users churn]
• [Churn patterns by segment]
• [Quote evidence of churn drivers]
• Example: "Enterprise users cite lack of custom integrations; SMB users cite steep learning curve"

From Sales/CS Meetings:

• [Customer feedback on product stickiness]
• [Feature requests from churned customers]
• [Usage patterns that predict retention]
• Example: "Customers who invite 3+ team members in week 1 have 90% D30 retention"

From PRDs:

• [Past retention improvements and their impact]
• [Features designed to increase engagement]
• Example: "PRD-2024-04 added daily email digest, increased WAU by 12%"

Gaps in Knowledge

Based on internal context, we don't yet know:

• [Gap 1]: Specific behaviors that separate retained users from churned users
• [Gap 2]: Which segments retain best and why
• [Gap 3]: Impact of specific features on retention curves

Should I help analyze your retention data, or would you like to provide additional metrics first?

Step 1: Retention Diagnostic Questions

Instead of generic "track retention metrics," I'll ask:

Question 1: The Biggest Drop

"Between Day 1, Day 7, and Day 30, where do you lose the most users?"

This tells me whether the problem is immediate product issues (D1→D7) or habit formation (D7→D30).

Question 2: Retained vs Churned Behavior

"What specific actions do Day 30 retained users take in their first week that Day 7 churned users don't?"

This is the differentiating behavior, not your opinion of what matters.

Question 3: Churn Reasons

"From churn interviews or feedback, what are the top 3 reasons users stop using?"

This tells me whether it's product quality, insufficient value, or competition.

Question 4: Usage Pattern

"How often should active users return—daily, weekly, or monthly?"

This determines whether D7 or L7 retention is your right metric.

Question 5: Segment Differences

"Do different user segments (size, industry, use case) have different retention patterns?"

Enterprise vs SMB, solo vs team users often have very different retention curves.

Key Retention Metrics

Day 7 (D7) Retention

Definition: % of users active on Day 7 after signup

Why it matters: Early signal of product stickiness

Benchmarks:

• Consumer social: 40-60%
• Productivity tools: 30-50%
• B2B SaaS: 50-70%
• Marketplace: 20-40%

Formula:

D7 Retention = (Users active on Day 7) / (Users who signed up 7 days ago) × 100

Day 30 (D30) Retention

Definition: % of users active 30 days after signup

Why it matters: Indicates habit formation

Benchmarks:

• Consumer social: 25-40%
• Productivity tools: 20-35%
• B2B SaaS: 40-60%
• Marketplace: 15-30%

Formula:

D30 Retention = (Users active on Day 30) / (Users who signed up 30 days ago) × 100

L7 and L28 (Rolling Retention)

Definition: % of users active in a 7-day or 28-day window

Why better than D7/D30:

• Accounts for usage patterns (weekly tools, not daily)
• More forgiving for non-daily products
• Better for B2B products

L7 Formula:

L7 = (Users active at least once in Days 1-7) / (Total signups) × 100

L28 Formula:

L28 = (Users active at least once in Days 1-28) / (Total signups) × 100

Retention Curves

Three types of retention curves:

1. Flattening Curve (Good) ✅

• Retention drops initially, then flattens
• Indicates core user base forming
• Example: Facebook, Slack

2. Declining Curve (Bad) ❌

• Retention keeps dropping over time
• No product-market fit
• Example: Failed consumer apps

3. Smiling Curve (Best) ✅✅

• Retention drop