RecoverPay

Verdict

As described, "dunning for the billing platforms the big players cover poorly" has about a 4% chance of reaching $10k MRR by month 24 and about 1.5% of reaching $20k (judgment estimates). The premise is mostly wrong: most of those platforms aren't underserved. Paddle gives its recovery suite away free. Stripe includes Smart Retries and recovery emails in its Billing fee, and a dozen tools already sell on top of it. Shopify subscription apps are Churn Buster's and FlyCode's home ground. Lemon Squeezy offers no API to retry a charge and is being folded into Stripe.

One platform really is underserved: Razorpay, with UPI Autopay and card e-mandates in India. None of Churn Buster, Churnkey or Baremetrics lists it. Stripe won't retry India-issued cards. RBI rules create failure types that global tools don't model. Cut to that wedge, the odds become ~50–65% for $1k MRR, ~22–34% for $5k, ~11–20% for $10k and ~4–9% for $20k (~15 hrs/week vs full-time, with conditions). Not fundable, and it doesn't need to be. See 01.

The five things that matter

1. The category is proven for bootstrappers. Churn Buster has been bootstrapped since 2013 and made about $996K in 2020 with about 11 staff (Latka estimate). Churnkey is estimated at ~$1.7M in 2025 with 15 staff. Stunning has sold flat-fee dunning since 2012. Recovered revenue is a number a merchant can see, which makes this one of the few micro-SaaS categories that sells on ROI.
2. The platforms are the competition. Stripe says it recovered $8.2bn of failed payments in 2025, with 55% recovered on average, all included in the 0.7% Billing fee. Paddle Retain is free on Paddle Billing, and since Jan 2025 every Paddle Billing account gets 7 retries over 30 days by default. Razorpay advertises its own "Smart Payment Retries". Your product must beat a free default, and prove it.
3. Involuntary churn is about a third of all churn. Recurly's July 2026 network data shows involuntary churn at 1.06 of 3.22 points for SaaS (~33%) and 1.25 of 3.60 points across industries (~35%). It's higher in low-ticket B2C (education 1.69, digital media 1.59) and almost nil above $250 revenue per customer (0.18). Aim at low-ticket B2C subscriptions, not enterprise SaaS.
4. India's rules change the product. A 24-hour pre-debit notice is required before every recurring charge. Recurring card charges above ₹15,000 need extra authentication. Customers can cancel UPI mandates from their app. "Smart retry timing" matters less. Mandate health, WhatsApp outreach and a one-tap UPI pay link matter more. That's where you win.
5. The ceiling is low. Indian merchants pay rupees, and the number of Razorpay merchants billing ₹5 lakh or more a month in subscriptions is unknown (estimated in the low thousands). $20k MRR may mean a tenth of the reachable market. $5–10k is the realistic target, and $20k needs a second payment rail.

Feature sort (headline)

• Table stakes: retries, card-update emails, the recovery page. Every platform already has them.
• Differentiators (for the wedge): WhatsApp-first recovery with UPI pay links, mandate-health monitoring (India's version of pre-dunning), and a holdout-proven "we recovered ₹X" report.
• Distractions: cancel flows in v1, multi-platform coverage, global USD sales, human collections calls.
• Liabilities: inflated recovery claims, holding or routing money, unregistered SMS, raw API keys stored carelessly.

Most likely cause of death

No provable lift over the platform default. Razorpay retries anyway. If your report can't show incremental rupees against a holdout group, merchants cancel in month 3, and the partner listing won't save you. Early warning: median lift below 2× your fee by day 90.

Next 90 days

Do: a Razorpay API spike in week 1. Then 20 merchant calls, 5 paid leakage audits, and 10 paying merchants on a WhatsApp-first recovery v1 with a holdout by day 90. See 08. Stop: planning integrations for Paddle, Lemon Squeezy, Shopify and Stripe. They're taken.