5. Pre-Mortem: It's September 2028 and RecoverPay Is Dead
Three most likely causes of death, ranked by probability. They compound: #1 usually causes #3.
Cause #1: No provable lift over the platform default (probability ~40% of failure cases)
What happened. Razorpay already retried failed charges and emailed customers. RecoverPay added WhatsApp messages and a monthly report showing "₹1.8 lakh recovered". A finance head at one merchant compared it with the previous quarter and found that most of those payments would have come in anyway. There was no holdout group. She posted about it in a founders' WhatsApp group. Month-3 retention fell to 55%, and new merchants asked for proof nobody could give.
Early warning signs
- By day 90, median lift against the holdout is under 2× the fee.
- Merchants ask "how do you know this wouldn't have been paid anyway?" and you don't have a clean answer.
- The monthly report is opened by fewer than half of merchants.
- Cancellations cite "Razorpay does this already".
Fixable? Yes, if it's built in from day one. Run a holdout on every account and report only incremental rupees. If the lift is genuinely small, move the product to where the lift is largest: mandate-health prevention and UPI Autopay switching, not reminders.
Cause #2: The market was too small, or wouldn't pay (probability ~35%)
What happened. The first 20 merchants came from warm intros and liked the product. Then growth stalled. There turned out to be fewer Razorpay subscription merchants billing ₹5 lakh a month than hoped, and many smaller ones wouldn't pay ₹2,499 for "reminders". Sales cycles ran 4–8 weeks because founders wanted to see the audit and check with their tech team. At ~15 hrs/week, the founder closed 2–3 merchants a month and lost 1–2. MRR plateaued at about $2.5k.
Early warning signs
- After 40 merchant calls, fewer than 15 qualify (₹5 lakh+ a month in subscriptions, on Razorpay).
- Audit-to-paid conversion is under 20%.
- Most revenue sits on the Starter plan.
- Churn above 4% a month from merchants shutting down or moving to Chargebee.
Fixable? Partly. You can widen the market with a second rail (Cashfree, PayU, Indian SaaS on Stripe) and sell through billing agencies and Razorpay partners. You can't make Indian merchants pay US prices. If the qualifying count is under ~1,000, cap your ambition at $3–5k MRR and run it as a profitable small business.
Cause #3: Razorpay shipped it, or cut you off (probability ~25%)
What happened. In 2027 Razorpay added WhatsApp payment reminders and a "recovery analytics" tab to Subscriptions, free for all merchants. Separately, a change to API access (a partner-programme requirement, or a new scope for payment-link creation) broke onboarding for three weeks. Merchants on Starter and Growth cancelled first. The product was left serving a few Scale merchants.
Early warning signs
- Razorpay product announcements that mention dunning, recovery, WhatsApp reminders or mandate management.
- Your partner-programme application stalls, or Razorpay asks you to stop using raw API keys.
- Paddle has already shown the pattern: it made retries default for everyone in Jan 2025 and bundles Retain free. Assume Razorpay will copy it eventually.
Fixable? Partly. You can't stop Razorpay. You can:
- be multi-gateway by month 15, since a gateway won't recover payments for its competitors' merchants;
- own the data, with the quarterly failure report and cross-merchant benchmarks that no single gateway publishes;
- become Razorpay's partner rather than its gap, with a listed app and a shared-revenue offer, so building it themselves looks less attractive.
Honourable mentions (less likely, still real)
- Security incident. A leaked merchant API key lets someone create payment links or read customer data. One incident ends trust in a small, talkative market.
- Messaging channel shutdown. WhatsApp downgrades your number's quality rating, or a DLT template is rejected, and recoveries stop overnight. Use per-merchant sender IDs and keep templates strictly transactional.
- The founder can't sell. This is a sales-led product for its first 50 customers. A builder who avoids calls will have a great product and ten customers.
- Rules change. RBI e-mandate rules have changed several times since 2021. A new framework (reported for April 2026, unverified) could change pre-debit or retry rules. Budget time to re-read the rules each quarter.