Hey IH
Building EarningsScores — an AI that scores every earnings report in real-time so retail investors can get institutional-grade analysis without reading 50-page filings. Here is the honest update from month 3.
Numbers:
- Users: 6 (up from 3 last month)
- MRR: $0 (still in free beta)
- Monthly costs: ~$200 (Vercel, API calls, infrastructure)
- Tickers scored this earnings season: 52 (27 beats, 11 misses)
What worked:
- Letting the product be the marketing. Every scored earnings report becomes an indexed page. People searching for specific tickers find EarningsScores organically. This is slow but it compounds every quarter.
- Building in public on Indie Hackers. The conversations here have been more useful than any amount of Twitter posting. Real founders giving real feedback beats broadcasting into the void.
What did not work:
- Finance Twitter outreach. Spent hours crafting threads and engaging with fintwit accounts. Result: maybe 2 signups from weeks of effort. The audience there wants hot takes, not tools.
- Asking friends and family to "check it out." They say they will. They do not. This is not a distribution channel.
What I shipped:
- Scored 52 earnings reports this season with the 6-layer AI scoring framework
- Goldman Sachs scored 72/100 today (beat by 7.2%, 16 buy ratings)
- Dashboard now shows real-time scoring during market hours with beat/miss/hold/buy/avoid signals
Biggest lesson:
The "feels productive vs actually compounds" distinction changed how I think about every hour I spend. Writing a Twitter thread feels productive but produces a one-time result. Adding a scored ticker page to my dashboard feels boring but keeps generating visits for months. I have started asking myself: if I disappeared for a month, would anything I did today still be generating results? If yes, it was leverage work. If no, it was maintenance.
Next month goal:
Get 3 users who are NOT friends or family. Real strangers who found the product on their own and came back a second time.
Would love feedback from anyone in fintech or building AI products. What am I missing? What would make you pay $19.99/mo for something like this?
Check it out: earningsscores.com
The 'compounding vs. maintenance' mindset is the only way to survive the early stages. If you’re asking what would make someone pay $19.99/mo, the answer is usually speed. Investors pay for the time you save them from reading 50-page filings.
Since your goal is to find 3 'real strangers' to pay you, you should jump into the Validation Arena (tokyolore.com).
It’s a 30-day $19 entry sprint where founders compete to turn their products into real revenue.
The prize pool is at $0 right now, and since your entry fee is basically one month of your target subscription, it’s a perfect stress test for your pricing.
The grand prize is a trip to Tokyo! 🏆