Building a new revenue
stream with no playbook

2026 is the year of prediction markets. Polymarket and Kalshi are processing billions per week. The category went mainstream overnight, and demand for "help me make money" was everywhere I looked.

I expected that to be easy to capitalize on.

OddsJam's moat is data access. Real-time sportsbook odds, exchange liquidity, pricing history across dozens of venues — getting all of it fast and correctly requires years of infrastructure and relationships competitors can't easily replicate. Own the data, and the recommendation layer is relatively straightforward.

Headwinds from all directions

Prediction markets broke that model entirely. Polymarket and Kalshi publish their full order book for free. When the data is public, hobbyists flood in — API wrappers, leaderboard trackers, Discord bots, all of it free, all of it low quality. The tools became a commodity overnight.

The silver lining: when everyone has the same data, the winners are whoever can do something opinionated with it. Have an opinion, make a call, stand behind it. That's the one thing none of the free tools would do.

BUT opinionated algorithms and recommendations aren't our strong suit either (it's data ingestion). We ended up stalling. let someone else prove the category, find an opportunity that fit our existing strengths better.

Prediction markets kept coming up internally. The volume was too big and the attention too concentrated to ignore. But no one could answer the hard questions: what can we reasonably build, and can we even sell it? Without that answer, everyone waited for evidence that required someone to go create it first.

Weeks passed. The window felt like it was closing. We had to move.

A solution wasn't going to appear on its own.

In an attempt to get the ball rolling, I ran two tracks in parallel. The first was the narrative. I studied what the commodity tools were doing and what they weren't: following the leaderboard doesn't work because leaderboard leaders are often arbitragers, market makers, or bots. The real levers were filtering those out, sizing based on each whale's confidence, and tracking exits alongside entries. I wrote that pitch before the algorithm existed — because a customer won't understand the algorithm anyway. If the narrative made someone immediately want access, that was evidence we could sell it.

The second track was actually building it — getting directional results from a rough algorithm, enough to show it worked in principle.

Prove you can sell it before you fully build it.

Nobody is going to tell you exactly what they need to see to be convinced. If that was the case product and business in general would be very easy. You have to show up with both: proof the thing can work, and proof people will pay for it. One without the other leaves too much open.

Putting the narrative and the directional results in front of people at the same time was what changed the room. Engineering could see it was buildable. The marketing team could see how to sell it. That combination created its own momentum.

We won the conviction to commit full resources. It became the biggest first-day launch in OddsJam history.

When there's an obvious opportunity and no obvious path to it, waiting for one to appear is its own decision. The move is to create the evidence yourself — on both sides — and let the response tell you whether it's worth going all in.