I don't think it's fraud, or anything like it. After all, there is something so-far sustainable here, and it's not the companies; it's the pattern (which is not quite new), and fraud is never sustainable. I have a theory about why we see this pattern, and it starts with how those companies are financed.
VCs fund many promising startups, often flooding them with money that can support the company for a long time, while maintaining a very active public image that perpetuates the sense that the company is successful. Not only is the actual business value hidden, VCs encourage the companies to not try to turn a profit, but to grow very, very quickly. This is, perhaps, what obscures the actual value the most.
Now, why do VCs do that? Because they hope for a good ROI, and some of them do quite well (the entire portfolio taken together, of course). Some say that VC's true desired goal for a company is an IPO, but IPOs are rare, and, I think a good IPO (for anyone who's not Facebook) is only about 10x that of a good acquisition. So I believe, that it is the acquisitions (that outnumber IPOs more than 10x) that really drive the VC investments, and, in turn, the whole industry.
But how can acquisitions be the bread-and-butter of the industry if so many of them fail (for the acquirer, that is)? Because, on the whole, acquisitions are still much cheaper for the acquirers than funding their own technology - or market - research. Instead of throwing a lot of money on large, money-hungry research departments, Big Tech would rather let a ton of entrepreneurs and VCs fight it out, and award the winners handsomely. The price they pay is far less than what they would have had to invest doing independent research.
So, who loses? I'm not sure anyone does. Entrepreneurs get the independence, excitement, and the possibility of huge payoffs of a winner-takes-all market; VCs - well, some of them - do alright, and Big Tech saves a ton of money on thousands of employees they don't need to directly employ and manage. Oh, and bloggers and industry insiders get a lot of juicy gossip and cautionary tales.
The one remaining question is, how come so many companies seem very promising, get good indications from the market, and then slowly (or quickly) declines. I don't have a good answer, but I do have a hypothesis. Web companies mostly compete in a global market. That means that in order for them to succeed, all relevant consumers must learn about them, and must learn about them quickly (fast growth, right?). But this is just not possible, because the average consumer can only keep in mind a bounded (and rather small) set of vendors. So, not only is, say, Groupon competing with uhmm, I dunno, Amazon, maybe for the purchase of some items, it is also competing with Zynga over my time. And not only that, it's even competing with Salesforce because there are only so many products I can even remember to use on a regular basis. And when new startups are funded, they are encouraged to very quickly get global attention, and - out with old, in with the new - novelty seeking consumers forget about yesterday's big thing.
So all of these companies are competing with one another for attention, so the question is, how many fast-growth, global companies can even prosper at the same time?