I'm the cofounder of Infura.io and we were building a 2-sided marketplace of API providers for blockchain infra. We bootstrapped it by doing exactly what you said: focus on one-side first. We focused on the provider side but we had that luxury. We had demand on the customer side already with our existing SaaS product. What we wanted to do was evolve our product to serve our existing customer base with a marketplace instead of solely by our product team. This is because of the evolution in the blockchain space of people interacting with dozens of APIs instead of 8 years ago when it was just our main Ethereum API. Serving the existing customers with a marketplace of providers made a ton of sense and so that is the direction we took the product. Over 2 years we grew it to a network of over 40 providers serving several dozen blockchain APIs.

For your idea I’d start narrower. One route or type of sender/package. Pick a high volume popular route and see if you can bootstrap the 2 sides just on that route alone and then expand. Then manually match people for the first 50-100 transactions. It takes some manual work to get that flywheel going