Jeremy Le-Tran
NYC
← Back
OddsJam, 2024–2025

Defining a
product with
no reference.

Building a premium product in a category with no precedent.

Role
Product Lead
Team
4 eng / 1 design
Timeline
10 weeks to launch
Outcome
$2M ARR projected

This isn't an aggregation product or a copy from another industry. Prediction-market APIs are free. Everything on top is technically a wrapper. Profitable traders keep their strategies private. The industry assumption was that a sellable paid tool couldn't exist here.

Building it required generating a new exploit from scratch: studying demand on social media, talking to top prediction traders, talking directly to Polymarket and Kalshi leadership, generating metrics from scratch, and testing the story on sales and customer success before committing. Conviction came from prototyping — both the algorithm MVP and the landing page — not from A/B testing alone. You can't A/B test a product that doesn't exist yet.

Diagram 01
The research funnel
01
Social demand
crypto/PM twitter
02
Direct convos
traders, Polymarket, Kalshi
03
Internal test
sales, CS, marketing
04
Algo MVP
prove directional profit
05
Landing MVP
offer + objection handling
Output
Conviction
to commit full resources. Algorithm + narrative co-designed, not sequenced.
Diagram 02
Why other options failed
Option A
Cross-market arbitrage
  • Fuzzy matching across naming conventions = weeks of work
  • Non-sports markets resolve too slowly
  • Less attractive than our existing arb product
→ skip
Option B
Market making
  • Special API access required
  • Manual is too slow
  • Full auto = turning OJ into a hedge fund
→ skip
Option C — build
Opinionated whale-tailing
  • Naive version = commodity (given away on Twitter)
  • Opinionated version = sellable
  • Filtered, confidence-weighted, time-aware
→ ship
Diagram 03
Objection handling
Objection
“Can't I just follow the biggest whales on the Polymarket leaderboard?”
01
Filter followable whales
Exclude market-makers and pod operators. Exclude non-scalable strategies.
02
Weight by confidence
Historical ROI, bet-sizing consistency, contract type, order-book state, entry vs. fair.
03
Flag time-sensitive
Entries above fair value after the whale bought are unreliable. Skip them.
“We took the decision-making out of it. Opinionated, not aggregated.”

The strategic work was eliminating options, not choosing one.

Arbitrage and market-making both could make money; neither could be productized and sold at $500. The landing page was part of the prototype, not a deliverable after. MVP'd the algorithm with one engineer to prove directional profit. The moment that worked, we committed full resources.

Outcome: Biggest 2-day launch in OddsJam history. Copycat designs appeared days later — proof we defined a new category. $2M ARR projected. First premium-tier ($500) consumer product.