FeedWise

Three most likely causes of death, ranked by probability. They compound: #1 makes #3 worse, because a founder with 8 customers stops watching the sources closely.

---

Cause #1: Nobody paid for a list they could copy (probability ~45% of failure cases)

What happened. The dataset was right, the API was clean, and the changelog was accurate. Developers loved the free package: 600 GitHub stars and 900 free keys. But when asked to pay ₹1,999 a month, most engineering leads said the same thing: "We update our config once a year, and NACH handles holiday shifts anyway." By month 9 there were 11 paying customers and ~$450 MRR. The free package had become the product.

Early warning signs

• In the first 30 buyer calls, fewer than 4 can name a real incident from a wrong holiday in the last 12 months.
• Design partners pay for the file but never set up the webhook.
• Free-key usage grows, and paid conversions stay under 3% of active free keys.
• Buyers ask "can you just send us the CSV once a year?"
• The best leads come from global holiday searches, not India bank or settlement terms.

Fixable? Only by changing the buyer, not the product. Try brokers and wealth apps (settlement dates cost more when wrong), or firms abroad with Indian payment runs (USD budgets). If neither pays by day 90, the free sources really are good enough. Stop, and move to another idea in the portfolio.

---

Cause #2: The data was wrong once, at the worst moment (probability ~30%)

What happened. A state declared a holiday on a Saturday evening for the Monday, during a festival week. The founder was travelling. The alert went out 19 hours late. Two lending customers ran collections calls and sent due-date messages to borrowers in a state where banks were shut. One posted about it in a fintech engineering Slack group. Three customers churned in a month, and no new Enterprise deal closed after that.

Early warning signs

• Median time from official notice to published change creeps above 4 hours.
• Any change is found by a customer before you found it.
• The scraper for one source (a state gazette, an exchange circular page) silently breaks and nobody notices for a week.
• No named backup for weekends and festival days.
• The checking routine is skipped on busy days.

Fixable? Yes, if it's designed in from day 1.

• Watch every source every 30 minutes, with alerts to two people.
• Track "silence" as a failure: if a source hasn't changed or responded in its normal window, alert.
• Publish a public changelog with "found at" and "published at" times, so speed is visible.
• Pay a backup for weekends and festival weeks from the first ₹25k of MRR.
• Terms that cap liability and say "verify with RBI".

---

Cause #3: The ceiling arrived early (probability ~25%)

What happened. The kill test passed. The product reached $2.8k MRR at month 14 with regulated lenders and a few brokers. Then growth stopped. The reachable buyers were already customers or had built it themselves. The next customers were seed-stage fintechs that died within a year, so churn rose to 4%. Adding cut-off times took 3 months and added 6 customers. At month 24 it was at $3.4k MRR against a goal of $10k, and the founder had spent 24 months on it.

Early warning signs

• Month 12 below $3k MRR full-time, or below $1.5k at ~15 hrs/week.
• More than 30% of new MRR from startups under 2 years old.
• New customers a month falls below 5 for 3 months in a row.
• Every sales call needs a custom feature.

Fixable? Partly. The ceiling is structural: one country's developer market for one kind of date. The honest options:

• Accept it. $3–5k MRR at ~90% margin, low churn and ~5 hours a week of upkeep is a real asset. At the Acquire.com median of 3.9x profit, it's worth roughly $125–210k (before paying for your own time).
• Add one adjacent line with the same shape (filing due dates that move by notification, for accounting SaaS). Test demand before building it.
• Don't add countries or datasets hoping to break the ceiling. That's cause #2 waiting to happen.

---

Honourable mentions (less likely, still real)

• A free source closes the gap. A contributor adds an RBI bank-holiday category to python-holidays, or RBI publishes an API through a government API platform. The list becomes free. Only freshness and history remain as reasons to pay.
• A funded player adds India depth. TradingHours.com or Calendarific adds RBI centre-level holidays. They'd win global buyers, but local freshness and rupee pricing could still hold Indian ones.
• Customer death. Indian fintech consolidation takes out 20% of your MRR in a year.
• Burnout from the pager. Daily source-watching with no end date is draining at ~15 hrs/week. A missed backup plan turns into cause #2.