Charles Ponzi arbitraged postage coupons (and resorted to selling ahead when that failed to turn profit). Strict application of your definition makes Ponzi's scheme not a Ponzi scheme.
As skeptical as I am about Groupon, it may work, in that it is not impossible by four orders of magnitude. Maybe they've got some brilliant pivot planned. Ponzi could never have made good using his putative business plan without fundamentally rewriting mathematics. But the real point isn't even that his plan was impossible, but that he wasn't executing the putative plan. Groupon is definitely soliciting businesses and issuing coupons.
[1]: http://www.mijiki.com/what-is-a-ponzi-scheme.html - I mean this to show my work here, I'm not sure that's a trustworthy site.
I'm not convinced that this follows from what we know about the numbers. It looks like one could just as easily make a case for the numbers indicating that the moment Groupon stops spending like crazy, their revenue will drop like a rock. What if they must cold-call dozens or hundreds of businesses in order to generate a sale? What if there isn't a way to sustain that without those aggressive sales tactics and the huge staff of sales people? What if they can't keep consumer interest without spending millions on marketing? What if there is a saturation point where people are simply bored of Groupon deals? (I unsubscribed after about 4 days, but I'm not much of a coupon kinda guy, so I don't know that I'm the ideal data point.)
I'm not saying it can't be true, I'm just saying I don't see that it is obvious that the hypothesis that there is a super-profitable business (or, at least profitable enough to justify their current valuation) underneath all the growth expenses is true.
I'm curious what you base this assertion on?