Why Monthly NPS is Already Too Late
Most founders obsess over quarterly NPS reports. By then, the damage is real.
Your best users notice friction first. They churn quietly. Then your score tanks. By the time you see it, 2-3% of your paying base has already left.
YC's Michael Seibel tells founders: measure NPS weekly. Not perfectly. Just consistently.
Here's the math: if you survey 100-150 active users per week, your sample size is statistically valid (±8 points margin of error). You'll spot a 5-point dip within two weeks. That's when you still have leverage to fix it.
The Weekly Workflow
Monday: Survey 100-150 users
Use Typeform, Qualtrics, or a simple Google Form. Ask one question: "How likely are you to recommend us to a colleague? 0-10."
Target: active paying users from the last 14 days. Exclude free tier (they inflate scores).
Include one open field: "What's the main reason for your score?"
Push rate: aim for 30%+ response. Offer a $5 Starbucks code if needed. Spend $200/month on incentives. It's cheap R&D.
Wednesday: Segment and Interpret
Calculate NPS the standard way:
(% Promoters 9-10) - (% Detractors 0-6) = NPS score
But don't stop there. Segment by:
- Signup cohort (month 1 vs. month 6 users)
- Feature usage (power users vs. casual)
- Plan tier (starter vs. enterprise)
- Use case (use case A vs. use case B)
If your overall NPS is 45 but "month 1" users score 10, you have an onboarding problem. You don't have a product problem.
This is the non-obvious part: NPS variance across segments beats absolute score. A 30-point spread between cohorts tells you exactly where to aim.
Thursday: Talk to Detractors
Pull the 5-8 detractors (0-6 score). Schedule 10-minute calls with 2-3 of them. Same day.
Don't ask: "How can we improve?"
Ask: "You gave us a 4. What would need to change for you to give us an 8?"
Listening matters. Writing matters more. Capture the exact word they use.
Scott Belsky's "Messy Middle" principle applies here: the gap between vision and outcome is where founders hide. Detractors show you the gap.
Friday: Decision and Commit
If a cohort dropped 3+ points week-over-week:
- You have a bug or a UX flaw.
- Ship a fix or a hotfix by Monday.
- Or announce a rollback/change within 48 hours.
If no movement (±2 points), continue. You're stable.
If NPS rises 2-3 points two weeks in a row, you're doing something right. Double down on it.
Real Signal, Not Theater
Many Indian founders (Toppr, Unacademy, Oyo in early days) used NPS this way. Weekly surveys. Founder-led detractor calls. Sunday night shipping.
They found product-market fit faster because they treated NPS like a stethoscope, not a trophy.
Comparison:
Theater NPS: "Our Q3 NPS is 52. Let's celebrate."
Signal NPS: "Segment B dropped from 38 to 35. We broke something in the Thursday deploy. Shipping a revert tonight."
The Math of Early Detection
Assume you have 1,000 paying users. Churn is 3% monthly (industry average for B2B SaaS in India).
If you measure quarterly (every 12 weeks):
You detect churn after 30+ users have already left. Recovery takes 8+ weeks. By then, you've lost 60+ users.
If you measure weekly:
You detect a 5-point NPS dip (your leading indicator of churn) in 2 weeks. You ship a fix in week 3. You recover in week 5. Total loss: 12-15 users.
That's the difference between a fixable problem and a death spiral.
Detractor Segmentation Framework
Category detractors by root cause:
1. Feature gap: "You don't have X." → Product roadmap decision.
2. Usability: "I can't figure out Y." → UX/onboarding fix.
3. Performance: "It's slow." → Engineering sprint.
4. Support: "No one replied to my ticket." → Process fix.
5. Pricing: "Too expensive." → Business model decision.
If 60% of detractors cite feature gap, build it. If 60% cite usability, don't build anything—fix onboarding first.
This is how you avoid the "we built the wrong thing" trap.
Implementation Checklist
- [ ] Set up weekly survey (Monday mornings)
- [ ] Decide on sample: active paying users only
- [ ] Create a simple spreadsheet: date, score, segment, reason
- [ ] Book 2-3 detractor calls (Thursday)
- [ ] Define action threshold: 3+ point dip = ship/rollback decision by Friday
- [ ] Share weekly NPS with team every Monday
- [ ] Track NPS by cohort, not just aggregate
The Non-Obvious Insight
NPS doesn't predict churn for you. Your action on NPS predicts churn.
Founders who measure weekly but ignore detractors see NPS flatten while churn rises. Founders who measure weekly AND call detractors AND ship fixes see both metrics move together.
The score is useless. The system is everything.
Actionable Next Step
Start this week. Survey 50 users today. Get responses by tomorrow. Pick 2 detractors and call them Friday. Ship something—anything—by Monday based on what you hear.
You don't need perfection. You need momentum. Weekly NPS cycles build it.