Thin Markets
Thin markets are markets that theoretically could exist, but are not active, or not fully functional, due to practical roadblocks and impediments. For example:
- Complex Items of Exchange – When the items in question are complex or nuanced, it is hard for the two parties to be sure exactly what they are giving or receiving. To conduct the exchange, they will have to share a LOT of knowledge, raising trust issues and demanding a lot of effort.
- Human Brokers are Deemed Essential – If you see a possible area of exchange that is not automated, but instead is dominated by human brokers with bulging contact lists, it is probably a thin market.
- High Value or High Risk Items – Buying a package of toilet paper off of Amazon is a low risk exchange. Even if it doesn’t work out, the losses are minimal. Selling 50,000 tons of high grade malting barley is a much higher risk endeavor. It’s not something that will likely be done with one click.
- Geographic and Cultural Barriers – If the sender and receiver are far apart, they may know little about one another. If their cultures or languages differ, they may know even less.
- Regulatory Barriers – If there are complicated regulatory hoops to jump through, the parties may not understand them. That is where brokers and middlemen have traditionally operated
- Complex Logistics – If the logistics to fulfill the exchange are complex, difficult, or changeable, then even if you could automate it, it might be hard to build an economic volume of transactions.
These are only a few examples of things that can get in the way of a fast easy exchange. If the roadblocks are primarily informational, there is a chance that an AI-based market might overcome them. If they are physical or regulatory, that may resist automation for much longer.
There are two types of thin markets: Markets that exist today, but operate inefficiently due to their structural challenges, and markets that can’t exist, but might emerge if there were a way to overcome the roadblocks. There is a third, more speculative, category if innovations were to inspire new and different forms of productive exchange. Collectively, these categories describe hundreds, probably thousands, of situations globally where participants might want to engage in exchange, but can’t manage the mechanics.
The possibilities for emergent new markets further muddies attempts to answer the question. Even if we could inventory all of the existing, inefficient markets, who knows how many new markets might be possible. Regardless, the numbers must be big, especially if one considers markets that are widely dispersed geographically.
Thin markets, as defined by Nobel laureate Alvin Roth, are characterized by scattered buyers and sellers with difficulty finding each other and low probability of random matching due to variety in needs and offerings. The economic research reveals these markets suffer from fundamental inefficiencies that AI-driven platforms might directly address.
Cosolvent
Cosolvent is designed to be a hosting framework for a market-matching site. Most of the operational logic will be carried out by various LLMs + RAG that are fine tuned to the task. Selecting those assets, as well as defining the traditional data schemas is the market developer’s responsibillity. Cosolvent provides the scaffolding so the developer can start site building faster, easier and more cheaply.
No. Thin markets, at least initially, are not good environments for conventional eCommerce. For example, it would be hard to imagine that an Indonesian brewery would buy 100,000 tons (4 or 5 containers) of malting barley, for delivery from Canada to Jakarta with a button click. A shipment that size will have shipping quotes, legal review, insurance bids, banking letters of credit, etc. etc.
Moreover, almost every thin market will have unique requirements for defining and handling the items that are exchanged. It seems unlikely that there will be enough commonality to aggregate different markets onto one platform.
It’s possible that a thin market automation tool like Cosolvent might eventually build enough users and usage that the market would become thick enough to switch to an eCommerce model. But not yet.
Cosolvent is totally free, under the MIT license. It is copyrighted, so you can’t simply claim it as your own, but you can clone the Github repo and make any changes you want, for as long as you want, and owe DeeperPoint nothing. Bear in mind, however, that DeeperPoint intends to make steady improvements and additions to the core version and you may want to think carefully about how or how far you want to break away in your own direction.