Overview
The product was a Rust-based MEV arbitrage engine: a narrow, technical sale to buyers who are hard to list and harder to reach. The buyers were, however, describing their own problem in public: Upwork does not stop clients from naming their company inside a job post, and many do. Tidalstead monitored MEV-related posts, extracted the company names, and opened conversations off-platform. After five to six weeks of passive monitoring, one of those conversations included a buyer acquiring an earlier version of the engine for about $20k.
A buyer list that does not exist
There is no database of firms that need an MEV arbitrage engine. Standard firmographic targeting has nothing to filter on, because the need is not a company attribute.
The marketplace’s economics work against you
Playing Upwork the intended way means spending connects, competing on price against a queue of bidders, and paying platform fees on whatever closes, for a product that is not a freelance gig.
Signal buried in noise, continuously
MEV-related posts appear irregularly and are only useful while they are fresh, so catching them meant watching all the time rather than checking occasionally.
Tidalstead’s Approach
We always start from first principles thinking. A job post is not a job. It is a company declaring, publicly and in detail, that it has a problem and a budget. Read that way, the marketplace is an intent database that happens to have a bidding system attached, and nothing requires you to use the bidding system.
Read the demand feed as an intent database
An automator monitored MEV-related posts continuously, so no window of freshness was missed.
Recover the account behind the post
The company name, named by the client in the post itself, was extracted and the account enriched via Lusha, turning an anonymous listing into a known buyer.
Move the conversation off-platform
Outreach went by direct email, as an account-based conversation rather than a bid. Zero connects spent, zero platform fees.
Results
- ~$20k sale, sourced entirely from this loop
- 5-6 weeks of passive monitoring before it landed
- $0 spent on connects or platform fees
- A marketplace demand feed permanently converted into an account-based motion
Conclusion
The signal was free, public, and sitting inside a system designed to charge for access to it. Reading job posts as declared intent, then leaving the marketplace for the actual conversation, produced a five-figure sale on no platform spend at all.