Hate Twitter? Then you're probably loving a new, buzz-generating study — released last week by the Texas market-research firm Pear Analytics — which found that the vast majority of Twitter messages, a/k/a tweets, are pretty much worthless. Over a two-week period, the study's authors randomly gathered 2000 tweets and then placed them into six categories of varying worth. The big winner: "pointless babble" (e.g., tweets with info like "I am eating a sandwich now"), which garnered slightly more than 40 percent of the total. "Conversational" tweets, or IM-type communications between users, came in a close second, with 37.5 percent. "News," meanwhile, was the loser of the bunch, comprising less than four percent of the sample.
But how telling are these numbers, really? Let's start with the study's methodology — which, regrettably, bears no resemblance to the way people use Twitter in real life. Twitter's great strength is that it allows users to determine, with minimal effort, both who and what they want to follow. If you do this — and pretty much every Twitterer does — you'll never encounter the terrifying info-tsunami that represents Pear Analytics' conceptual starting point.
What's more, Twitter's build-it-yourself quality makes the study's value-laden categories nearly useless. If you're following a given topic (updates about a particular neighborhood, say) or a given individual (a friend, a politician, a favorite athlete), you're going to value the ensuing tweets in a way that others don't. In other words, one Twitterer's "pointless babble" is another's "news" — a nuance the study somehow managed to miss.
Just one more thing: in a blog post introducing the study's results, Pear founder and CEO Ryan Kelly closed with an eager plug: "Since Twitter is still loaded with lots of babbling that not many of us have time for, you should check out the Twitter filter, Philtro." Unmentioned by nearly everyone who covered the study — with the laudable exception of the UK's Register and thecustomercollective.com — is the fact that Philtro's founder and CEO, Paul Singh, seems to be the same Paul Singh who serves as Pear Analytics' business-intelligence expert. (Both went to George Mason, both were founder and COO of the mobile-messaging company UpSolv, and both like to be called "Mr. Metrics.") Shaky reasoning and a blatant conflict of interest? Pointless babble, indeed.
