I learned quite a bit experimenting with Magento and a multi-vendor extension, as well as trying to sell a few items myself. It is certainly tough to get one of these started.
I do still feel that a niche-oriented marketplace immediately surrounded by a number of valuable non-marketplace services and content oriented toward the niche collectively have the ability to gain significant market share.
A correctly executed platform is only one aspect of the approach, but a significant one. The others are a heavy dose of networking, an unwavering belief in the concept, a willingness to use the tools you're trying to convince others to use, a dash of creativity, and an acceptance that while some things might fail they should be taken as an exercise in learning... and applied going forward. Hopefully, once I figure out the correct formula I'll be in a position to share more.
I'm unsure, but recall eBay at one point had bought Magento, I believe, which may have since changed. Separately, I've come across a few job postings with Accenture requiring familiarity with the platform so I suspect it's measurable.
I had tried Magento along with uMarketplace Suite. The latter was very well suited for it's intended use-case: multiple vendors selling through a centrally organized marketplace along with centrally managed fulfillment workflows, tax considerations, etc. But, my conclusion was that for my project's purposes something more bottoms-up rather than top-down makes sense. I may have mentioned elsewhere, but Magento - while very capable - was incredibly complex. Their default database had something northward of a hundred of database tables.
I believe Lamps.com is built upon Magento, and probably uMarketplace Suite as well. An example of how well a platform can be built on each - but together they're not intrinsically a social marketplace :).