agentsclimarketplace

Retention analyzer

Skill reatlat/fullstory-claude-plugin/skills/retention-analyzer

User retention analysis — N-day retention, cohort stickiness, returning vs new user patterns, and churn signals. Use when asking about retention rates, user stickiness, churn analysis, or "are users coming back?"From its SKILL.md

Install
npx -y skills add reatlat/fullstory-claude-plugin --skill retention-analyzer

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

4.6 KB, as published. Nobody here has run it

Retention Analyzer

Measure user retention — how many users come back, how often, and who's at risk of churning.

When to Use

  • "What's our 7-day retention rate?"
  • "How many new users from last month are still active?"
  • "Are enterprise users more sticky than free users?"
  • "Show me users who haven't returned in 30 days"
  • "What's the churn rate for users who signed up in Q2?"
  • "Compare retention between users who completed onboarding vs those who didn't"

Mental Model

Retention is about returning users. There are three common lenses:

  1. N-day retention: What percentage of users who were active on Day 0 came back on Day N? (e.g., Day 1, Day 7, Day 30)
  2. Cohort retention: Of users who signed up/started in a period, what percentage are still active N days later?
  3. Stickiness: How many days per week/month do active users engage? (DAU/MAU ratio)

Fullstory doesn't have a native retention report, so you build it from segments and metrics.

Workflow

Step 1: Define the retention window

Clarify what the user wants:

  • N-day return: "Are users active on Day 7 and Day 30?"
  • Ongoing engagement: "How many sessions per week do active users have?"
  • Churn detection: "Which users haven't returned in 30+ days?"

Step 2: Build a retention cohort

For N-day retention from signup:

fullstory:build_segment("users with first_seen between July 1 and July 7") → cohort_segment

For ongoing activity:

fullstory:build_segment("users with last_seen in last 7 days") → active_users
fullstory:build_segment("users with last_seen before last 30 days") → churned_users

Step 3: Measure return rate

Build a metric that counts users who were active in both the cohort window AND the return window:

fullstory:build_metric(
  query="users who were active between July 1-7 AND active between Aug 1-7",
  output_type="single_number"
)

Compare to the total cohort size to get the retention rate: "1,200 of 5,000 July Week 1 users were active in August (24% 30-day retention)."

Step 4: Engagement depth

For stickiness (how often active users engage):

fullstory:build_metric(
  query="sessions per user in last 7 days",
  output_type="top_n"
)
→ "45% of active users had 1-2 sessions, 30% had 3-5, 15% had 6-10, 10% had 10+"

Step 5: Segment retention

Using the comparisons skill approach, break retention down by cohort:

By plan:
Enterprise: 52% 30-day retention
Free:        18% 30-day retention

By onboarding:
Completed onboarding: 41% 30-day retention
Skipped onboarding:    9% 30-day retention ← BIG GAP

Step 6: Churn signals

Identify users at risk of churning:

fullstory:build_segment("users with last_seen 14-30 days ago AND total_sessions > 5")
→ these are previously engaged users who've gone quiet

For these users, check:

  • Did they hit errors or frustrations in their last sessions?
  • Did their session frequency drop before they stopped?
  • What was the last thing they did?

Use user-journey to trace a sample of churned users and look for common patterns.

Step 7: Present

Format the retention analysis:

## Retention — July 2024 Cohort (5,000 new users)

### Return Rates
- Day 1: 68% (3,400)
- Day 7: 41% (2,050)
- Day 30: 24% (1,200)

### By Segment (30-day)
- Enterprise: 52% | Free: 18%
- Completed onboarding: 41% | Skipped onboarding: 9%

### Churn Risk
- 890 users haven't returned in 14-30 days (previously active)
- Common last action: viewed /pricing, did not convert

### Recommendation
Onboarding completion is the biggest lever — users who finish onboarding are 4.5x more likely to stick. The pricing page is where users go before churning.

Guidelines

  • Fullstory doesn't have a built-in retention report. Be transparent: "I'm building this from segments and metrics — it's accurate but takes a few extra calls. For dedicated retention analytics, consider a tool like Amplitude or Mixpanel."
  • Day 0 = user's first day (from first_seen). Day 1 = the next day, Day 7 = a week later, etc.
  • For B2B products, retention is typically measured in weeks or months, not days. Adjust the window.
  • "Active" usually means "had at least one session in the window." Clarify with the user if they define it differently.
  • Segment-level retention differences are often the most actionable insight — "enterprise users retain at 52%, free at 18%" tells you where to focus.

Keep looking

Skills are one crate of 614,194. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.