I Thought the World Had Supply and Demand Figured Out. I Was Wrong.
By Spencer Connaughton
Railroad Tycoon
When I was a teenager, my favorite game was Railroad Tycoon. Not because of the trains, but because of the economics. The game displayed a heatmap showing where goods cost little (abundant supply, low demand) and where they cost a lot (constrained supply, high demand). Your job as the player was to connect the two, running trains from surplus to scarcity.
I assumed someone had solved supply and demand discovery in the real world long ago. Surely, since the 1800s, great thinkers and industrialists had built the economic systems to match supply with demand at scale. It would explain the abundance we’re privileged to experience in the early 21st century. I thought our leading citizens had surely built the Railroad Tycoon heatmap, and the world was running a continuous program of optimization and connection.
I was wrong.
The Reality
While working at Emerson Collective, I started to realize the world might be more disorganized than I thought. Our portfolio companies were building physical products, and all needed standard input goods to do it. Manufacturing teams buying metals, fasteners, and generic chemicals had to hire full time staff to run RFQ processes. Green builders would sometimes give up on RFQs entirely and run to the hardware store to pay retail prices for wood, nails, and glue.
During my time at Google X, I experienced this problem first hand sourcing components for devices we were building. Even with digital catalogs and all the e-sourcing tools we had at Google, I would still find myself on the phone with suppliers asking "How much can I buy, and at what price?” and keeping track in a spreadsheet.
The real “ah-ha” moment came in a customer discovery meeting. My colleague Ray Benato and I were exploring what would become Mattera. We sat down with a senior executive at one of the nation's largest recycling companies. First, we asked questions, listened, and gathered unbiased information about his operations. Then came the fun part, where we reveal what we’re working on. In a typical meeting, this would wow and inspire our new partner to join us on our journey to the future. Ray cleared his throat and said:
"What would you say if I told you that instead of getting 10 types of plastic out of your recycling process, you could get 100?"
The executive didn't even pause.
"What would I do with 100 new outputs?"
Surprised, Ray said he'd sell them, right alongside his other outputs. The executive shook his head:
"I've got a team of 10 people, and all day long they're texting and emailing, trying to find buyers for the 10 types of plastic we already produce, and I can barely keep them organized. I’m not sure how I’d manage 100 merchandising managers, and I'd have no way to know if they were getting good prices."
I noticed nearly all of our customers and development partners doing the same thing: buyers reaching out to multiple suppliers, every supplier quoting a slightly different price, and all negotiations happening over text messages, emails, and phone calls. All humans, all manual.
The cost of connecting buyer and seller, one text message at a time, created a transaction cost barrier that made entire categories of commerce uneconomical. This is when I learned that the bottleneck isn’t process technology or logistics. The bottleneck is the conversation.
Conversational Trading
I explained this pattern to my friend and mentor Chris Yeh, who gave it a name: conversational trading.
I decided to see how widespread this business practice was. I called more than 100 people across dozens of verticals: metals, plastics, chemicals, agricultural products, building materials. Every single one described conversational trading. Some recalled efforts to introduce exchange trading in their verticals, which would inevitably run into the cold start problem and fail. Many spent millions on e-sourcing and procure-to-pay platforms with built-in reverse auction tools, which would prove too complicated, and require too much buy-in from counterparties. They eventually went back to text and email.
Everyone assumed that other industries had this figured out, and were eagerly awaiting the day market transparency and efficiency would come to their own industry as well.
According to the U.S. Economic Census, $11.3 trillion of “wholesale goods” trade annually in the US. By my estimate, somewhere between 30% and 100% of those goods trade conversationally.
Every failed exchange made the same fundamental mistake: they tried to replace the conversation. They asked people to change their behavior all at once, and to abandon the tools they know (SMS, WhatsApp, email) and learn something new.
That doesn't work. Not for the veteran broker who has texted deals for 20 years. Not for the buyer who can reach every supplier from their phone. Not for anyone who has built a career and relationships on bilateral negotiation.
Conclusion
I couldn’t stop thinking about how to bring an entire market into the future without asking anyone to disrupt their workflows or trade at a disadvantage. Our solution would have to:
Meet people where they already operate: in their text messages and in their inboxes.
Build market structure around the conversation instead of trying to replace it.
Massively incentivize the brokers, the gatekeepers of these markets, instead of trying to cut them out.
And we figured out how to build it. In my next piece, I'll explain what conversational trading actually looks like, and how we make it easier for everyone.
