What the Heck Is Twitter?

It’s not a Google killer, and it’s not a Facebook killer.

For you and me, Twitter is a fun way to procrastinate. But for Silicon Valley’s chattering classes, the microblogging company has emerged as something much more—the next Google, the next Facebook, or maybe some unbeatable combination of the two. “It’s time to start thinking of Twitter as a search engine,” TechCrunch’s Michael Arrington declared this week, joining a chorus that includes John Battelle, the San Jose Mercury News, and some of Twitter’s venture funders. By collecting millions of people’s immediate thoughts, Twitter is building the Web’s best database of “real time” information, these people argue. And that collection might be very valuable—when people want to know what’s going on in the world right now, they’ll increasingly check Twitter, not Google.

Some of the great expectations surrounding Twitter can be discounted as classic Silicon Valley hype; in an otherwise hopeless time, it stands as one of the few things the tech world is still optimistic about. Twitter is also extremely simple—so simple that it’s often tempting to describe it as something more than it is. Perhaps that’s why, in trying to capture Twitter’s potential, boosters compare it to known successes—search engines and social networks. The trouble is, neither comparison makes much sense.

Let’s start with the Twitter-beats-Google theory. For proponents, two recent events stand out: last year’s terrorist attacks in Mumbai and the water landing of US Airways Flight 1549 in New York. Both times, people on the scene began Twittering what they were seeing and hearing, and Twitter’s search engine became one of the first places on the Web to carry the news. Indeed, the first picture of Flight 1549 in the water came via Twitter: Janis Krums, a businessman who was on a nearby ferry, snapped a shot and posted it along with a note, “There’s a plane in the Hudson. I’m on the ferry going to pick up the people. Crazy.”

In a recent blog post, John Borthwick, one of Twitter’s investors, explained the value of such a service. Imagine that you’re waiting in line for coffee, and you overhear some people talking about a plane landing on the Hudson. When you go back to your desk, you search Google—and come up with nothing. “The day of the incident there was nothing on the topic to be found on Google,” Borthwick writes. “The same holds for any topical issues—lipstick on a pig?—for real-time questions, real-time branding analysis, tracking a new product launch. On pretty much any subject if you want to know what’s happening now, will come up with a superior result set.”

Todd Chaffee, another Twitter backer, told that Twitter might even be able to answer personalized search queries—stuff that Google doesn’t know about. “You put a question out to the global mind, and it comes back,” he said. “Millions of people are contributing to the knowledge base. The engine is alive. You get feedback in real time from people, not just documents.” John Battelle describes this scenario: Say you were interested in buying a certain digital camera, the Canon EOS. Instead of going to Google to look for information about the camera, you’d send your query to everyone on Twitter. “Does the Canon EOS take good pics?” If Twitter could build a service that analyzes and sorts through responses in real time, showing you the best few responses on top, “those results could be truly game changing,” Battelle writes.

But there are two problems with this idea. First, Twitter’s search engine, at the moment, is extremely crude. It sorts results only chronologically, so the best answers are easily swamped by silly ones and people repeating silly ones. Of course, Twitter could improve its search engine. But so could Google. Twitter, remember, is a public database. Everything on Twitter is on the Web and can be analyzed by any other firm—in fact, Twitter’s own search engine began as a startup unaffiliated with Twitter. (Twitter bought that company, Summize, last summer.)

Google already maintains many specialized search engines—a blog search engine, a book search engine, a news search engine, a video search engine, a Usenet search engine, and others. If searching the “real-time Web”—as some people describe Twitter—ever became a popular pastime, it would be a trivial matter for Google to start indexing Twitter and then displaying tweets alongside its main results. And if it did this, Google would clearly have the upper hand: Search for “Canon EOS” in Google, and you’d get Web pages and news stories, plus the most recent, relevant tweets. Search for the same thing in Twitter, and you’d get just the tweets. Which would you choose?

True, it’s possible to think of scenarios where real-time searches would yield better results than what can be found on Google. If you were visiting Boston and looking for a good place to grab a bite near Harvard, you might be better off asking your Twitter network to recommend a place than looking for something on the Web. Indeed, people do this all the time on Twitter. But compared with everything we look for online, these personalized, just-in-time searches are in the minority. Think of everything you Googled yesterday—how many of those searches required real-time results that were generated by others just because you asked for them? In fact, when you do ask Google for something really peculiar, you often find that someone has already asked that question at Ask Metafilter or Yahoo Answers. Even Battelle’s example doesn’t really work: Why does anyone need to know what people are saying about the Canon EOS right now? Wouldn’t you care how someone reviewed the camera yesterday or last week? Is that information outdated? Of course not—and often, it’s much deeper than information you’ll find on Twitter. That’s the key point: For most searches, information that’s merely timely, rather than immediate, is good enough.

That gets to a more fundamental question: How often does anyone need “real time” information on the Web? Yes, in the first few minutes after a plane has landed in the Hudson, Twitter might be the best place to find news. But its advantage is short-lived. By the time TV helicopters begin circling overhead, Twitter becomes decidedly slower and noisier than other media—now you’re much better off going to CNN, where, in addition to a live picture, you’ll find comments from authorities, structured interviews with eyewitnesses, and other old-fashioned journalistic information.

Shortly after the Mumbai attacks began, I opened up one Web page to Twitter’s #mumbai stream and another to live coverage from the Indian TV news network IBN Live. The TV coverage was fraught with all the problems that usually plague breaking-news reports—spotty, speculative information; contradictory witness statements; and endless repetition. But the Twitter stream was even worse. It was incoherent, telling no narrative—just a continuously updated jumble of facts, pseudo-facts, unfounded assertions, opinions, rants, condolences, and, most of all, repetition of information that I’d heard on TV minutes before. Even Google surpassed Twitter: Very quickly, the search engine began to index news stories, live blogs, and the Wikipedia page for the attack. Searching for “Mumbai” in Google gave you an in-depth, though still timely, picture of what was going on. This happens for every news event; it is simply not true, as Borthwick asserts, that there was nothing on Google about Flight 1549 on the day of the accident. * Google had a coherent story within hours; Twitter had a messy story within minutes. News junkies love the messy, fast story, but lots of us aren’t news junkies. Lots of us are OK waiting an hour for coherence.

So if Twitter is not a threat to Google, is it a threat to Facebook? That case is more straightforward. As Facebook founder Mark Zuckerberg wrote this week, one of the social-networking site’s main goals is to let people “share immediate experiences with one another: a thought, a status, a photo, a note, a feeling.” But Twitter is a faster, purer, less rule-bound way to share all that stuff. With Twitter, you just sign up and type—you don’t need to build a profile or worry about whom to friend and reject. (By default, Twitter lets anyone follow your updates; on Facebook, generally two people must agree that they’re friends before they can share notes.) And Twitter is fast, displaying what people in your network are up to as soon as they press send. Facebook, until now, has been a bit slower—the site shows you new updates from your friends only every few minutes or so, not instantly.

In a sign that it recognizes the appeal of Twitter, Facebook announced that it will soon change its home page to allow for immediate updates. You’ll be able to see what your friends are doing right now, just as you can on Twitter. It’s a good move, but probably not revolutionary. Again, I think people overestimate the appeal of “real time” data. There are some instances in which real-time interactions will improve Facebook—when all of your friends are watching the State of the Union address, for example, you might want to watch everyone’s status update change as the speech progresses. But will we be much more satisfied to get our friends’ status updates immediately rather than have to wait 10 minutes? Some people may be overjoyed; most likely won’t notice the difference.

It wasn’t too long ago that Twitter was the butt of Silicon Valley jokes. It makes no money, and until recently its chief contribution to the culture was its much-displayed service-outage logo, the “fail whale.” Twitter is growing extremely quickly—traffic leaped by more than 700 percent in 2008—and lately it has captured the attention of just about everyone in politics and the media. But despite its growth, it’s still a niche product: gets about 6 million American visitors a month, according to traffic-monitoring firms. Facebook beats it by a factor of 10 and Google by a factor 20; even the Huffington Post bests Twitter.

That’s not to say Twitter doesn’t have a great future ahead of it. As more people join it and as we get better tools to analyze the cacophony of conversations taking place there, we’ll find it increasingly useful when news breaks or when we need instant feedback from our friends. What’s unclear now is how such a service will make any money—and how it can transform itself in order for that to happen. It’s possible that Twitter could do very well—but probably not as a social network, and probably not as a search engine.

Correction, March 9, 2009: This piece incorrectly referred to an “attack” on US Airways Flight 1549. The plane was forced to make a water landing when it struck a flock of geese. (Return to the corrected sentence.)