You could certainly make money on transactions via a market built into the group system. You can imagine baseball card or antiques or furniture collective groups running their own selling & auctioning (the market could enable a few different selling methods). Buyers could flow from one group market to the next with one account, and feedback in the market would cross groups which would amplify the value proposition of the whole system further.

If you add the market on top of the group social features the build time will risk becoming incredible, unfortunately. You're looking at building a potentially enormous platform. Maybe with three to five people and a stripped down set of features. Could also make it a non-profit, open source it and run a hosted version of it as the parental core (more or less what WordPress does), scaling up the number of contributors to accelerate development. The mission (debasing some of Facebook's power) and ethos (non-profit) might be enough to draw people willing to contribute.

The transaction fees in the market (drop below eBay's most common ~10% fee, aim for 3% or 5% perhaps) would of course go to fund the non-profit's efforts to build it all out, avoiding donations and advertising permanently.

I think the critical factor would be: how stripped down can you make it to get started (while still having something people will want to use), to keep build time as low as reasonably possible.