The Two-Bucket Framework
Michael Seibel taught this at Y Combinator. Your growth is stuck. Before you pivot, answer one question: Are users arriving but not staying, or are no users arriving at all?
These need opposite fixes. Conflating them kills startups.
Bucket 1: Users Arrive but Leave (Churn Problem)
You have 500 active users. Last month, 600. Month before, 700.
But here's the tell: new users do sign up. Daily. You're not a ghost town.
This is a product problem. Maybe the onboarding is confusing. Maybe users hit a wall at day-14 when they realize your value prop doesn't match reality. Maybe competitors solve it better. Doesn't matter why—your product isn't keeping people.
YC founders here need to persist on product, not pivot away. Superhuman spent 2+ years obsessing over email client retention. Their churn was 2% monthly. They survived because they fixed why users left, not because they pivoted to "faster email" or "AI email" every quarter.
The action: Install proper analytics. Measure churn by cohort. Find the day retention drops. Build for that moment.
In India, see Dunzo's early mistake: they pivoted from "student delivery" to "grocery + hyperlocal" without nailing retention in either. Users arrived, used once, left. They pivoted instead of diagnosing why.
Bucket 2: No Users Arrive (Acquisition Problem)
You've built something. Small team loves it. Retention is solid—70% of users who try it stay.
But nobody's trying it. 5 signups last month. 3 the month before.
This isn't a product problem. This is a go-to-market problem. Your funnel is broken. Channel-market fit is missing. Maybe your target audience is too small. Maybe your messaging is off. Maybe you're selling enterprise software to consumers or vice versa.
Persevere on product. Rebuild acquisition. Hire a sales person if B2B. Try a different channel if consumer. Test a new cohort.
Doesn't require a pivot—requires iteration on sales strategy.
Yet Indian founders here do pivot. They built an employee wellness app. Nobody signed up. So they pivoted to student wellness. Still nobody. Then to hospital wellness. Three pivots, same underlying problem: nobody in their TAM actually feels their pain acutely enough to pay.
That's a positioning or TAM problem, not a product problem. Pivot makes sense only if you discover a different audience does feel the pain acutely.
The Diagnosis Framework
Asking the right question first:
Q: Do you have engaged users?
- YES → Go to next question.
- NO → You have both problems. Likely a TAM or positioning issue. Small pivot may be warranted. But test first.
Q: Is your engaged-user base growing?
- YES → Pure acquisition problem. Hire a head of growth. Test new channels. Scale what works. Persist.
- NO → Retention problem. Cohort analysis. Why do users leave? Fix that before acquiring more. Iterate product.
Q: What's your Monthly Churn Rate?
- Consumer apps: Below 5% = healthy. Above 10% = urgent product work. 5-10% = monitor closely.
- B2B SaaS: Below 3% = you can scale. Above 5% = product problem before sales expansion.
Scott Belsky's "Messy Middle" Addition
Belsky argues that startups die in the middle—after product-market fit signals arrive but before true scale. Most founders here pivot prematurely.
The signal: You have some users who actively use, but growth is slow and inconsistent.
This isn't time to pivot. This is the Messy Middle. You need to:
1. Deepen engagement with existing users (product).
2. Systematize acquisition (operations).
3. Survive 12-18 months on this dual focus.
Founders mistake "slow growth" for "wrong product." They're usually wrong. Olacabs had slow growth for 2 years. Not because taxis don't exist. Because building trust and density in a city takes time.
India-Specific Friction
Tier 2+ founders often see churn confused with "users not being ready." Not the same thing.
Small Town USA doesn't have Dunzo, but users arriving and not returning = your service quality failed, not that the idea is wrong.
Second: multi-city rollout masks real churn. You expand to Delhi. Growth looks good. But your Bangalore cohort is dying 40% month-over-month. You don't see it until you slice by geography.
Way to diagnose: Measure month-over-month retention by cohort, by city, by user segment. Not just overall.
Decision Framework Summary
If retention ≥ 70% (any engaged segment) → Persist. Build better distribution.
If retention ≤ 50% across cohorts → Pivot is justified. Product doesn't match claim.
If retention 50-70% → Messy Middle. Survive it. Don't confuse survival for failure.
If you have zero users → Diagnostic pivot toward different positioning or audience is acceptable. But only after testing.
The Non-Obvious Insight
Most Indian startup subreddits and pitch decks claim retention without measuring it properly. They count "installed" or "created account" as retention. Real retention: did they come back unprompted 30 days later?
Fake retention metrics hide whether you have an acquisition or product problem. You think you're pivoting because growth is slow. Actually, you're acquiring users into a leaky bucket. Fixing the bucket first might be boring. It's also usually right.
Action: Next 48 Hours
1. Pull your last 4 months of weekly cohort retention data.
2. Identify your healthiest cohort (retention ≥ 70%).
3. If it exists: problem is acquisition. Hire or test new channels.
4. If it doesn't exist: product-market fit signal is missing. Diagnose which segment sticks longest and why.
5. Do not pivot until you've answered this.
Persistence and pivoting aren't opposites. Diagnosing correctly lets you do the right one.