How We Search With The Twitter “Help Engine”

Is Twitter a search engine or not? There’s been plenty of discussion and debate about this recently. I’d say yes, in a way. It’s clear to anyone who watches a twitterstream that people put out questions to Twitter similar to how they use search engines. But if anecdotal examples aren’t enough, a survey I conducted last week confirms that Twitter is used as a search resource.

If you’re scratching your head about the phrase “help engine” in the headline, my companion piece, The Rise Of Help Engines: Twitter & Aardvark, goes into more depth about the new class of search engines I’m calling help engines and why.

Before diving into the survey numbers, let me start out with one of the comments I received about the survey, which ran for several days:

I just want to mention that this poll is totally unscientific and thus the results are meaningless. It is likely totally biased toward those that use Twitter frequently and obviously biased toward those that use your site. If your aim is to get some anecdotal data, that’s fine, but please don’t try to draw any conclusions from this poll.

You don’t want to get me going about the number of “scientific” polls I’ve read over the years that are nonetheless meaningless for a variety of reasons. I’ll save the longer debate of the “science” behind many polls for another time. But the comment deserves some attention.

My poll cannot tell you what everyone does on Twitter with any certainty, in terms of search. Moreover, I’m not certain how anyone could get a “random” sample to produce such a poll. Those with many followers are not the same as those with a few. Do you calculate the percentage of highly followed users and ensure they answer in the same proportion as those who take a survey? And how do you “random dial” these people? What percentage of people are Twitter novices versus those with experience, which might also influence any answers about how they use Twitter for search?

No, this survey doesn’t provide a perfect picture of how people search via Twitter. But I think it’s a good first step beyond the anecdotes that people report. It provides a few preliminary hard numbers to put behind all that commentary about Twitter as search.

It’s also important to understand this survey is NOT about Twitter Search. That’s a separate part of Twitter, where people can explicitly do a deep web search against past tweets to find information. Instead, this survey was about how people use Twitter itself — their network of followers — to ask for help directly, especially when in the past, they might have first turned to a search engine.

Finally, when I write that Twitter is a search engine, or being used by some who see it that way, I think it’s important to understand that I don’t mean this is Twitter’s primary purpose — nor that Twitter as a micro-blogging service sees itself as a search engine. I simply think this is a by-product of the service.