/pulse

Retention Curve Analysis Reference

Purpose: Deep-dive retention curve shape classification, Power User Curve, Quick Ratio, and stickiness diagnostics. Anchored to a16z, Reforge, and Sarah Tavel frameworks. Produces actionable retention diagnostics with SQL and alert thresholds.

Scope Boundary

• pulse retention: Retention curve shape (L / smile / flat), power-user band, Quick Ratio, DAU/MAU (this document).
• pulse cohort (elsewhere): General cohort design and churn measurement. Default entry.
• pulse activation (elsewhere): Activation rate and aha-moment. Activation precedes retention in the funnel.
• Growth (elsewhere): Retention strategy (win-back, habit formation). retention owns measurement; Growth owns intervention design.

Workflow

SCOPE     →  confirm product type (B2B SaaS / B2C / mobile / marketplace)
          →  pick window (90/180/365-day minimum)
          →  define "active" event (per NSM)

QUERY     →  build cohort matrix (signup month × months-since-signup)
          →  compute retention % per cell

CLASSIFY  →  shape curve (L / smile / flat)
          →  identify stable plateau (>= month 3)

OVERLAY   →  Power User Curve (MAU engagement frequency band)
          →  Quick Ratio (growth / churn)
          →  DAU/MAU stickiness

DIAGNOSE  →  map shape to pathology (see table below)
          →  prioritize remediation: activation vs engagement vs resurrection

ALERT     →  define drift thresholds for cohort-over-cohort comparison
          →  handoff to Growth for intervention design

Curve Shape Classification

Retention %
 100% |                              (all curves start at 100%)
      |\.
      | \..
      |  \...
      |   \.....                     L-shape (BROKEN)
      |    \........                  → falls to 0-10% and keeps falling
      |     \............
      |
 100% |\.
      | \..
      |  \....
      |   \.....                     FLAT (STABLE, low)
      |    \......_______            → stabilizes near 15-40%
      |
 100% |\.
      | \..
      |  \...              __         SMILE (HEALTHY)
      |   \...__________///           → stabilizes, then rises (resurrection)
      |      stable band

B2B SaaS Benchmarks (2025 update)

• Month-1 logo retention: 46.9% (avg), ≥70% (healthy), ≥90% (elite)
• Month-12 logo retention: ≥80% (healthy), ≥95% (elite enterprise)
• NRR (Net Revenue Retention) — 2025 reset: private-SaaS median 101-102%; top performers 104-106%; SMB segment often 90-105% (below 100% increasingly common); enterprise 115-125%; best-in-class enterprise software 120-150%; growing companies with NRR ≥100% grew ~2x faster YoY than peers below 100% in 2025. See revenue-analytics.md for source-cited cuts.

B2C/Consumer Benchmarks

• Day-1: 25-40% (avg mobile)
• Day-7: 10-20%
• Day-30: 3-8%
• Elite social/utility: 25%+ at Day-30

Power User Curve (a16z)

Beyond aggregate retention, classify users by engagement frequency within MAU. Strong products show a concentration of users on the right-hand side (high engagement days per month).

% of MAU
  │
  │███
  │███             █              CLIFF (weak)
  │███             █              → Most MAUs engaged <5 days
  │███████        █               → Heavy left-skew
  │███████████████
  └───────────────────────
    1  5  10  15  20  25  30  days-active-in-month

  │                         ███
  │                        ████
  │                    █████████   SMILE (strong)
  │██████            ███           → Right-skewed MAU
  │██████████████████              → Healthy power-user band ≥21 days
  └───────────────────────
    1  5  10  15  20  25  30

Rule of thumb: teams should watch the L21+ band (users active 21+ days in 30). Elite consumer products have 30%+ of MAU in L21+.

Quick Ratio (Growth Velocity)

Quick Ratio = (New MRR + Expansion MRR) / (Churn MRR + Contraction MRR)

DAU / MAU Stickiness

Cross-check: if DAU/MAU is low but NPS and revenue are high, the product may be intentionally low-frequency (tax software, annual tools). Don't force daily usage where the job doesn't require it.

SQL Patterns

Cohort retention matrix (BigQuery / Snowflake)

WITH signups AS (
  SELECT
    user_id,
    DATE_TRUNC('month', signup_at) AS cohort_month
  FROM users
),
activity AS (
  SELECT
    user_id,
    DATE_TRUNC('month', event_at) AS active_month
  FROM events
  WHERE event_name = 'nsm_action_completed'
  GROUP BY 1, 2
)
SELECT
  s.cohort_month,
  DATE_DIFF('month', s.cohort_month, a.active_month) AS months_since_signup,
  COUNT(DISTINCT a.user_id) * 1.0 / cohort_size.total AS retention_rate
FROM signups s
JOIN activity a USING (user_id)
JOIN (
  SELECT cohort_month, COUNT(DISTINCT user_id) AS total
  FROM signups
  GROUP BY 1
) cohort_size USING (cohort_month)
GROUP BY 1, 2, cohort_size.total
ORDER BY 1, 2;

Power User band (L21+)

WITH monthly_activity AS (
  SELECT
    user_id,
    DATE_TRUNC('month', event_at) AS month,
    COUNT(DISTINCT DATE(event_at)) AS days_active
  FROM events
  WHERE event_name = 'nsm_action_completed'
  GROUP BY 1, 2
)
SELECT
  month,
  COUNT(CASE WHEN days_active >= 21 THEN user_id END) * 1.0
    / COUNT(*) AS l21_share
FROM monthly_activity
GROUP BY 1
ORDER BY 1;

Quick Ratio (monthly)

WITH mrr_movement AS (
  SELECT
    month,
    SUM(new_mrr) AS new_mrr,
    SUM(expansion_mrr) AS expansion_mrr,
    SUM(churn_mrr) AS churn_mrr,
    SUM(contraction_mrr) AS contraction_mrr
  FROM mrr_movements
  GROUP BY 1
)
SELECT
  month,
  (new_mrr + expansion_mrr) / NULLIF(churn_mrr + contraction_mrr, 0) AS quick_ratio
FROM mrr_movement
ORDER BY 1;

Drift Alerts (Cohort-over-Cohort)

Route CRITICAL alerts to Scout for investigation and Growth for intervention. Route HIGH/MEDIUM to product owner with cohort drill-down.

Deliverable Contract

When retention completes, emit:

• Cohort retention matrix (at least 90 days, preferably 180-365).
• Curve shape classification (L / flat / smile) with plateau % and benchmark comparison.
• Power User Curve with L21+ band percentage.
• Quick Ratio trend (last 6 months).
• DAU/MAU stickiness ratio.
• Cohort drift alerts with thresholds and severity.
• Diagnosis: activation gap vs engagement gap vs resurrection opportunity.
• Handoff targets: activation (activation gap), Growth (re-engagement intervention), Experiment (uplift validation), Scout (anomalous drop).

References

• a16z — "The Power User Curve: The Best Way to Understand Your Most Engaged Users"
• Sarah Tavel / Benchmark — "Consumer Retention Playbook"
• Reforge — Growth Models and retention measurement
• Brian Balfour — "Why Retention is the Single Most Important Growth Metric"
• Mixpanel — Retention Benchmarks Report 2024
• High Alpha — 2025 SaaS Benchmarks
• ChartMogul — The SaaS Retention Report: The New Normal
• RockingWeb — SaaS Metrics Benchmarks 2025 (2,000 companies synthesis)
• Benchmarkit — 2025 SaaS Performance Metrics