×

Can Twitter be Saved?

Even by Internet standards of hyper-growth, there has never been a phenomenon like Twitter. Less than a year and a half ago, Twitter hit 1 million users. It now has 44 million, a rate of expansion so rapid that if it could continue growing at that speed it would in another year and a half be used by everyone on Earth. It is impressive not just for the sheer number of users but for the share of mind it has carved out, from the national elections to its starring role in Iran's election protests. Twitter has become so ubiquitous so fast that it's almost impossible to imagine it disappearing.

But it can. The irony of Twitter is that even as it becomes more pervasive, it is in danger of very quickly becoming markedly less useful. Twitter is in danger of collapsing under its own weight. Not because of its problems keeping up with traffic—those are solvable—but because the volume of material that Twitter unleashes now puts impossible demands on its users' time and attention. The problem, in a nutshell, is information overload. The more Twitter grows and the more feeds Twitterers follow, the harder it gets to mine it for what is truly useful and engaging. Even as Twitter reaches a peak in the cultural cred cycle, it's time to start asking how it can be saved from itself.

Twitter
Source: Twitter
Twitter

I first started wondering about how much attention folks are paying to all their Twitter feeds about a month ago, when one of my stories was mentioned on the Google corporate Twitter feed. At the time it had close to a million followers (and is now at more than 1.3 million). I would have expected the link in the Google tweet to generate a substantial amount of traffic to the story. It did not. It's hard to pin down an exact number, but judging by The Big Money's logs of how readers came to the story, Google's tweet seems to have led about 100 to 200 people to click the link and read the full story.

I'm happy to have even a few additional readers, but this is a very, very small number. More important, the number of people who click a link in a tweet is probably the best proxy we have for a measure of how engaged folks are with Twitter. The small number of people clicking on that link seemed to me an early sign that we've reached a point of Twitter saturation, in which the result of more people following more feeds adds up not to more communication but more noise.

_____________________________________
More Web 2.0 Coverage from The Big Money:

_____________________________________

In the recent history of technology, we've often been told of the value of "network effects." Much of our experience of technology is with positive network effects and increasing returns as more people take advantage of it. An obvious example is e-mail; the more people use it to communicate, the more useful it is. Yet network effects can also be negative. A park that is popular becomes more vibrant and appealing. But a park that gets too popular is just crowded. Twitter's growth has been so rapid that it is clearly bumping up against the limits of its usefulness. It is not only increasingly full of noise, but the sheer volume of stuff coming through the Twitter fire hose renders even what was useful much harder to pick apart and make sense of.

A recent and heavily publicized study showed that the overwhelming majority of Twitter tweets come from just 10 percent of its users and that most Twitter users send out less than one tweet a week. Twitter skeptics seized on the study to say that the Twitter meme has been overblown; one memorable headline even took the study to mean that "Twitter is almost completely useless." This isn't really fair. For one thing, the Iranian protests proved without a doubt that in some situations Twitter really is stupendously useful. More generally, there's nothing wrong with signing up for Twitter and sending out nothing at all. Just because most people don't write books doesn't mean that books are useless. Just as most people are happy simply to read books, many Twitter users may prefer to follow a few prolific tweeters but rarely send out anything themselves.

Tax attention

The real issue with Twitter as it grows bigger is not how few people send out messages but how extremely prolific the top Twitterers are—and how profligate many users are in pressing the "follow" button. I am only an occasional Twitterer. Right now I follow 27 Twitter feeds. By Twitter standards, this is not by any means a big number, but I already find myself overwhelmed by the volume of messages these 27 people alone generate. Most of my followers follow many more than 27 streams; about 200 feeds is typical, and one follows 3,057. If I can't keep up with the mere two dozen or so people I follow (I've had to turn off the beeping notifications on TweetDeck to avoid distraction), I can't even imagine what it must be like for someone to "follow" 3,000-plus people. How much attention she can possibly pay to any of my tweets is very small.

If someone wants to follow me and 3,056 others and ignore us all, she has every right to do so. I am not offended. But it does raise the question of just how much engagement we can expect with Twitter as each user signs up to follow more feeds and each active Twitterer tries to tweet more and more often to break through the static.

In a useful experiment, Daniel Scocco, of DailyBlogTips (yes, to the few who may still be uninitiated, there are many blogs about blogging) asked his Twittering readers to track the proportion of their followers who clicked on the links in their tweets. The most common results fell in the range of about 0.5 percent—a link was clicked by about 1 in 200 people who got a tweet—though some outliers did better.

_____________________________________
More Web 2.0 Coverage from The Big Money:

_____________________________________

More interestingly, some of Scocco's readers sent in more detailed analyses, looking at the response to different types of tweets. In general, links that were closely related to the main subject of the feed did substantially better. One Twitterer, who runs a blog called Fermentarium about home brewing and winemaking and goes by Deege on Twitter, found that about 2.5 percent of his 4,429 followers clicked on links related to beer or wine, while when he went off-topic and tried a tweet on an unrelated subject, only 12 people followed the link.

These are very unscientific results, but they do seem to pass a basic reality check—they make sense. And between the DailyBlogTips experiment and my experience with Google's Twitter feed, I'm comfortable making some tentative generalizations. One is that corporate tweeting is inherently problematic. Corporations invariably believe that they take up a much bigger part of their customers' mental space than they actually do. As the novelty of Twitter wears off, the little attention that people do devote to most corporate feeds will decline even more; I suspect that in the near future many of them will seem as pointless as the inane White Castle home page of 1996.

For the feeds that do manage to clear the increasingly high hurdle of relevance, the implication of Scocco's little experiment is that the way to keep followers engaged is to narrow their focus and stay on topic. This is quite different from the strategy that many folks now seem to pursue on Twitter, which is to flood the zone with lots of tweets in the hope that simply sending out more messages is the key to getting noticed. Judging by the response to links, the advantages of merely sending out more stuff are quickly offset by the precipitous drop in interest as tweets become less on point.

The process of weeding out

Just as happened with the Web, the natural tendency of Twitter will be to eventually whittle down the number of feeds devoted to corporate proselytizing and self-indulgent moment-in-the-life catalogs. That process is already happening but not fast enough. What will accelerate it will be methods of automatic filtering, whether developed by Twitter or others, that will push what is most interesting to the top. Using programs such as TweetDeck, it's already possible to categorize and filter Twitter feeds.

That alone, though, is not enough to channel the Twitter torrent. There is just too much material in 200 feeds; a few people will make the effort to elaborately categorize it all, but most will not. Somehow, for Twitter to remain viable, this will have to be automated. That could be done by harnessing the power of groups, in the manner of Digg, with tools that will push up the tweets that other followers have found worthwhile. It could also be done (I suspect more effectively) by programs that let users mark the most important tweets and learn to look for and highlight similar material.

Whether people will be willing to leave this kind of selection to automatic agents is something that we'll have to see. In the case of e-mail, we are not usually willing to do that. We want our e-mail systems to get rid of obvious spam, but while the technology is there, very few folks seem willing to let a bot guide their e-mail reading. With Twitter things may be different; for most people, Twitter feeds fall into the "nice to keep up with" drawer and so are different from e-mail, in which we absolutely, positively don't want to miss anything important.

_____________________________________
More Web 2.0 Coverage from The Big Money:

_____________________________________

The disappearance of pointless feeds, the narrowing of truly useful ones, and the emergence of systems that will automate the process of judging what's important and highlight it are the baseline for what Twitter needs to avoid collapsing under its own weight. It's certainly possible, though, that all that won't be enough. Twitter is an extremely clever idea, and it will continue to exist in some form. But that form may be very different from what we see now. It may, for instance, eventually turn into a way of pushing information to Web sites and Facebook pages; we will then get the feeds in our ordinary use of the Web, without specifically choosing to become "followers" and accumulating an ever-growing collection of feeds.

Certainly the folks who run Twitter may have new tricks up their sleeves. Not having been clever enough to invent Twitter, I would be loath to try to advise the people who did on what they should do to manage Twitter overload. But a prediction I am willing to make is that just as the growth of Twitter has been so rapid that it caught almost everyone by surprise, the problem of containing and managing the overload will become paramount for Twitter sooner, more forcefully, and more suddenly than folks expect. A lot of people have already had the experience of seeing Twitter abruptly go from engaging to essential to all-consuming. But it's become a little bit like the map that Lewis Carroll once envisioned: having grown to be as big as the world itself, it ceases to be valuable as a map. The question now is whether Twitter can be tamed before it too has grown too world-encompassing to be useful, and the cacophony of signals turns into pure noise.