[line-retention-cohort] User Retention Curve by Cohort

[MarkusNeusinger]

Description

A line chart showing the percentage of retained users over time since signup, with separate curves for different cohorts. Similar to a survival curve but focused on product engagement rather than clinical events.

Key Visual Elements

• Time since signup on x-axis (days, weeks, or months)
• Retention rate (%) on y-axis (0-100%)
• Multiple lines, one per cohort, color-coded
• All curves start at 100% at time zero
• Characteristic exponential decay shape
• Legend with cohort labels and sizes

Applications

• Product analytics: comparing retention across cohorts
• A/B testing: measuring feature impact on retention
• Growth teams: evaluating onboarding improvements
• Investor reporting: demonstrating engagement trends

Domain

Product Analytics / Marketing

[claude]

✅ Spec ID: line-retention-cohort

Rationale: This is a valid and widely-used product analytics visualization. While related to survival-kaplan-meier, it differs significantly: retention curves use connected lines (not step functions), always start at 100%, don't involve censoring, and focus on cohort comparison for product engagement. No existing spec covers this use case.

Add label approved to begin generation.

[github-actions]

✅ Specification Ready: line-retention-cohort

PR: #4601

---

specification.md

line-retention-cohort: User Retention Curve by Cohort

Description

A line chart showing the percentage of retained users over time since signup, with separate curves for different cohorts. All curves start at 100% at time zero and typically exhibit exponential decay, revealing how well a product retains users over their lifecycle. By overlaying multiple cohorts, teams can visually compare whether retention is improving or degrading across signup periods.

Applications

• Product analytics teams comparing user retention across monthly or weekly signup cohorts to measure engagement trends
• A/B testing analysts evaluating how a new onboarding flow affects long-term user retention versus a control group
• Growth teams reporting retention improvements to stakeholders by showing cohort curves shifting upward over time

Data

• time_period (numeric) - Time since signup in consistent units (days, weeks, or months), starting at 0
• retention_rate (numeric) - Percentage of users still active, ranging from 0 to 100, starting at 100% for all cohorts
• cohort (categorical) - Cohort label identifying the signup period (e.g., "Jan 2025", "Feb 2025")
• cohort_size (numeric) - Number of users in each cohort, displayed in the legend
• Size: 6-12 time periods per cohort, 3-6 cohorts
• Example: Monthly signup cohorts tracked weekly for 12 weeks, showing active user percentages

Notes

• All curves must start at 100% at time zero
• Use distinct colors for each cohort with a legend showing cohort label and size (e.g., "Jan 2025 (n=1,245)")
• Y-axis should range from 0% to 100% with gridlines for easy reading
• X-axis represents time since signup, not calendar dates
• Consider using slightly decreasing opacity or thinner lines for older cohorts to emphasize recent ones
• A horizontal dashed reference line at a key retention threshold (e.g., 20%) can highlight target benchmarks

---

Next: Add approved label to merge this specification to main.

---

🤖 spec-create workflow

[github-actions]

🚀 Specification Merged!

Specification: line-retention-cohort
PR: #4601 (merged)

The specification is now in main and synced to PostgreSQL.

---

Next: Add generate:{library} labels to start generating implementations.

Available libraries: matplotlib, seaborn, plotly, bokeh, altair, plotnine, pygal, highcharts, letsplot

Or use workflow dispatch for bulk operations.

---

🤖 spec-create workflow