[byrneseyeview]

suddenly unpopular with both users and investors...

The first rule of bad journalism is to confuse the first and second derivatives. Groupon is enormously popular with users and investors--it's just not the most popular it's ever been. Yipit's fairly recent data indicate that while the daily deal market contracted slightly over the summer, but Groupon shrank more slowly than LivingSocial ( http://techcrunch.com/2011/08/25/yipits-daily-deal-report-gr... ).

In "Fooled by Randomness," Taleb points out that the more frequently you measure results, the more random they are. Second-by-second, Warren Buffett is losing money just about half the time.

Groupon obviously has problems, and LivingSocial is executing amazingly well, but it should be laughable to call them "unpopular" rather than "less massively popular than at their peak a couple months ago," in the same way that you wouldn't call someone "poor" because they moved down a spot in the Forbes 400.

[andrewljohnson]

I upvoted your comment, but I still think you might be wrong.

The journalist may be sensationalizing here, but a high-growth, pre-IPO start-up that is contracting, before it begins to exploit profits, is in trouble.

Whether it's the market or the company that is in trouble, or a mixture of the two, is debatable. But I do think "suddenly unpopular" conveys the gist of the story better than "contracting from its previous peak." The first tells me something is wrong in Groupon world (and there probably is), and the second tries to convince me the whole thing is a non-story.

[jonnathanson]

I think the word "unpopular," in the context of the article, is referring to people's perceptions of the company's business practices and brand. Not so much to the size or engagement of the userbase. I might be wrong, but that was my interpretation.

(It's "unpopular" right now because it's come under a lot of heat in the press; perhaps "embattled" would be a more accurate word).

[hammerdr]

And neither would be accurate. Most users, both businesses and customers, are wildly happy with group on.

[nopinsight]

Citations, please.

Without knowledge about their churn rates, especially in comparison with their competitors, it is not obvious either way. There are also users who still receive the daily emails because they did not bother to cancel the subscription, but rarely look at them.

[msellout]

I like Taleb, but I think you've misinterpreted him. High-frequency measurements of a random variable will reveal more about the probability distribution than will infrequent measurements. As far as I know financial time-series are scale-invariant. While infrequent measurement may take advantage of the law of large numbers in identifying the mean, frequent measurement will allow observation of both the mean and the variance of the distribution.