Zillow's plan was to offer to buy homes sight-unseen in targeted areas, using an algorithmic pricing scheme based almost entirely on location and limited home info. They hoped, I think, to gain substantially market power by owning a lot of the available inventory in the targeted areas.
Was it a stupid business plan? Yes, it was clearly a stupid businessplan. The plan invited an "adverse selection" effect wherein homeowners would accept their offers only if Zillow had offered substantially more than what the homeowners could reasonably expect to get on the open market.
I remember hearing a Marketplace radio segment on the program like a week after it launched, where it was completely obvious that this was going on.
It was a pretty clear business model, with a glaring and predictable flaw. Shareholders may have lost money, but I don't see how they were defrauded or misled.