My co-founder and I spent 2021 full-time building GreenChain: a B2B marketplace for a more sustainable food supply chain.

• The food supply chain comprises multiple layers: producer/grower/manufacturer → distributor/wholesaler/broker → foodservice/retail. The commerce is opaque, fragmented, inefficient, and primarily online. Our long-term vision was to become the digital platform for all food business buyers and sellers to operate and transact.
• To start, our GTM focused on small-medium restaurants and small-medium fresh-food distributors.

At the year’s end, we concluded that the idea, strategy, and model as envisioned would fail. We completely discarded all the work and used the lessons to start anew.

We ultimately did not believe the marketplace characteristics (product discovery, supplier discovery, and price comparison) were sufficiently valuable to the users. Most likely, we would have become a productivity tool for our users (small-medium restaurants and grocery retail) to order their supplies, track inventory, etc. And 15+ existing startups do that already (https://choco.com, https://rekki.com).

The marketplace dynamics -- connecting buyers to all of the suppliers that the service their geographic region -- felt less and less important:

1. If a fresh-food supplier is out-of-stock of a good (e.g., Atlantic Salmon), then likely all suppliers in the region would be struggling because it’s a supply-chain issue. Hence, finding alternative vendors in the event of a stock-out seemed less likely.
2. We believed that restaurants changed their menus too infrequently to find sufficient value in the “discovery” aspect of new products.
3. Price comparison was not that important because the price differences were minor amid market prices. And, most restaurants did not even know their exact prices because they would frequently change (minorly), and the restaurants wouldn’t bother to compare invoices.
4. Restaurants and suppliers are incredibly reliant on their sales rep relationships and contract sales. Doing one-off orders with a supplier was utterly foreign. We believed that we could automate all of the hurdles to make it possible, though we felt restaurants would not use it because they carefully evaluate new suppliers, try samples, etc. So restaurants might not even bother.
5. Deciding on a take-rate was very difficult because of the low-margin business. We believed we could justify a rate for replacing bringing new sales without a salesperson who usually would receive a sales commission. But, we worried we would not have the leverage for that amid competition for all startups building digital platforms for selling to their customers.