Peter Kafka

Recent Posts by Peter Kafka

Is Twittermania Running Face-First Into Quittermania?

weegee-crowdRemember all the way back, a couple weeks ago, when everyone was talking about Twitter and Oprah and Ashton Kutcher and the millions of people who were joining Twitter every week?

Turns out the majority of those new Twitterers won’t be back in May.

So says Nielsen Online, which estimates that 60 percent of Twitter’s users leave after a month. That makes sense on a gut level to me:¬†Twitter is easy to use, but it often takes a while to make sense, and if you’re not a professional self-promoter–or someone with a lot of friends who are already on Twitter–it may never make sense.

It’s worth noting here that Nielsen is likely overstating the churn because it is only measuring visits to the URL. The majority of Twitter use happens away from the site, on mobile phones and apps like Tweetdeck, and it’s theoretically possible to be an avid Twitterer but never visit after you sign up. I’ve asked the Twitter folks for their take on the stats and will update if they respond.

But let’s assume, for argument’s sake, that the Nielsen stats are correct, or close to being correct. Is that a problem? Obviously, every Web service attracts new users who never come back after they try it out, so churn in itself isn’t a problem. The question is the rate.

The good news is that Twitter’s 40 percent retention rate is higher than it used to be. Prior to the Oprah madness of this month, Twitter’s rate was closer to 30 percent, Nielsen says.

But the measurement company argues, via a fancy chart and equation, that 40 percent retention makes it mathematically impossible for Twitter to achieve significant penetration with Internet users. The simple version is that if Twitter loses three out of five users a month, its growth will be capped at about 10 percent of the audience. Fancy version below (click chart to enlarge):


And here’s how Twitter’s retention rate compares to that of Facebook and MySpace (again, note that Facebook users and MySpace users more or less have to visit the those sites to use them, so the numbers are likely slightly skewed):


So what if Twitter really is a service that appeals to no more than 10 percent of the Internet audience? Is that such bad thing? Not at all. That’s an awfully big number.

And “retention” may end up being the wrong metric to measure a service like Twitter, anyway. See this perceptive post by Andrew Chen (thanks to OMGPOP’s Dan Porter for the link).

But a lot of the Twitter sales pitch–to investors and would-be partners like Google (GOOG) and Microsoft (MSFT)–is contingent on the service’s eventual ubiquity. The appeal of Twitter’s real-time search capabilities, for instance, is less seductive if you’re only searching what a sliver of Internet users are Tweeting about. And knowing that growth is capped could make that impressive hockey stick chart a little less so.

[Image credit: Weegee via the International Center of Photography]