Primary data · sourced from public filings·700+ listed companies · India-first·
Open screener
ἀλήθεια · aletheiaAncient Greek for truth — literally “un-forgetting”: the act of revealing reality, not merely stating it
← All posts
Sector Thesis·4 min read·Week 26

The Cohort Analysis Every Founder Should Run Monthly

Cohort analysis reveals whether your product is genuinely improving or just acquiring more users. By grouping users by signup week and tracking their retention, you see the truth: improving product metrics or declining retention hiding behind growth vanity metrics. This is the single most important chart every founder should build.

ByAmit Tyagi·Fitoor Capital
Aletheia Insights · Weekly

Get 1 unfair insight every week from India's startup ecosystem.

Read by serious founders and investors. No fluff.

Why Vanity Metrics Lie to You

You launched a campaign. Downloads spiked 40%. The founding team celebrated. Month later, revenue flatlined.

You optimized ad spend. Now 500 signup requests daily. Still flatlined.

This is the startup trap: growth without retention is a leaky bucket. You're pouring water in the top. It just drains faster.

Cohort analysis forces you to see this. Not eventually. This week.

What Cohort Analysis Actually Reveals

Cohort analysis groups users by signup week (or month, for B2B). Then you track what percentage returns in Week 2, Week 4, Week 12.

Why weeks? Because it's granular enough to spot product changes quickly. Monthly data hides what happened in Week 1.

The magic: if Week 12 retention for January cohort is 8%, but February cohort hits 12%, your product got better. Not your marketing. Not luck. Your actual product.

This is the metric Y Combinator forces every batch company to obsess over. Michael Seibel's advice: "If retention is flat, everything else is noise."

How to Build the Chart in 90 Minutes

Step 1: Export user data (3 columns max)

Signup date, user ID, last active date.

If you don't have this, you're not ready for this analysis. Build it first.

Step 2: Create signup cohorts

Open Google Sheets. Column A: signup week (Week of Jan 1, Jan 8, Jan 15, etc.).

Column B onward: "Week 1 Retention," "Week 2 Retention," "Week 4 Retention," "Week 12 Retention."

Step 3: Calculate retention percentages

For January 1-7 cohort:
- Total signups: 150
- Returned in Week 2: 90 users
- Week 2 retention: 60%

Repeat for all weeks. Use COUNTIFS to automate.

Step 4: Plot the chart

Line chart. X-axis: signup weeks. Y-axis: retention percentage.

Each line represents one time period (Week 2, Week 4, Week 12).

If lines trend upward left-to-right, product is improving. If flat or declining, you have a problem money can't fix.

What Indian Founders Must Know

For B2C (Fintech, Food, Games):

Weekly cohorts matter. Day 1 retention matters most. If Day 1 retention is below 20%, the app itself is broken—not the marketing.

Example: Paytm's early cohorts tracked Day 1, Day 7, Day 30. When Day 7 stopped declining in March 2013, they knew the product was sticky. They scaled.

For B2B SaaS (HR Tech, Finance Tools, Analytics):

Week 1 retention is misleading. Most paid users don't touch the product until Week 2. Track Week 4 instead.

Example: A CRM startup in Bangalore noticed Week 1 retention was 50% but Week 4 was 85%. They stopped worrying about Day 1 engagement. Instead, they built better onboarding for Week 2-3. Week 4 retention climbed to 92%.

Margin mattered. Week 4 data changed their roadmap.

The Non-Obvious Insight: Your Best Cohort Is Your Blueprint

Most founders chase their worst cohort. Wrong move.

If March cohort retains 60% in Week 4 but April cohort is 45%, ask: What changed in March?

Maybe:
- Press coverage brought better customer type
- A specific feature shipped mid-February
- Onboarding flow changed
- Sales team was better trained

One of these caused the 15-point lift. Find it. Repeat it.

Scott Belsky calls this "messy middle archaeology"—digging into the chaos to find the signal.

What to Do When Retention Declines

If Week 4 retention drops 5+ points month-over-month, act fast.

First: Don't blame acquisition quality yet. Your product changed or broke something.

Second: Check if a feature shipped that week. Kill it or fix it immediately.

Third: If nothing shipped, ask your top 10 active users why others quit. Use Intercom or just call them. You'll find the answer in 2 hours.

Inaction is the real cost. Most founders ignore declining cohorts for 6 weeks. By then, they've onboarded 2,000 users to a degraded product.

The Monthly Ritual

First Tuesday of every month: Run this analysis.

Take 30 minutes. Share results with your team. One question: "Are we retaining better cohorts this month?"

If yes: Double down on what changed.

If no: Stop all marketing until you fix it.

This is not optional. This is how winners think.

One Clear Takeaway

Build your cohort chart by Friday. Plot 8 weeks of data. If the trend line is rising, keep building. If flat or falling, everything else is irrelevant. Fix retention first.

Amit Tyagi

Founder, AletheiaAI & GP, Fitoor Capital

Veteran of India's startup ecosystem. Writing about fundraising, investor psychology, and what it takes to build fundable startups in India.

Run a fundability check

India's only MRE-backed platform for founders and investors. Analyse your deck, find investors, and validate your raise strategy.

#cohort-analysis#retention-metrics#product-analytics#founder-metrics

Don’t miss the next one

One insight every week. No fluff.

Aletheia Insights · Weekly

One contrarian insight. Every week. No generic startup advice.

Join founders and investors building with better information.

The Cohort Analysis Every Founder Should Run Monthly · Aletheia Insights