I was browsing TechCrunch today and saw these predictions for tech trends in 2010. It’s a pretty good overview for conventional wisdom; one thing I noticed at number three was real time search, search engines that find the latest developments and most useful results based on what’s happening right now (i.e. if you searched for “bomb threat” a couple of days ago, the Time Square threat that was unfolding would be pushed to the top of the results because of its recent salience). Real time search is a very hot area right now, and several startups (TechCrunch name drops OneRiot, Collecta, Topsy, but they’re just the most successful tip of the real time iceberg) sprung up to capitalize on a flaw in Google and other traditional search engines. Their idea was that traditional search spit out results based on past measures (how many pages link to you, how many people visit your site) which inadequately respond to breaking events. Lately, both Google and Bing have both beefed up their real-time search (which is why one of the options on the side of a Google search is to look only at results from the last 24 hours), and, indeed, Twitter is reportedly now making a profit simply by letting Google and Bing pay them to index people’s tweets.
Unfortunately, real-time isn’t the problem with search, and doesn’t solve an issue that people really care about. Most searches have little to do with recent events: I might search for lyrics to a song, a menu for a restaurant, a place that sells Wii controllers etc. The times that I am searching for something recent (e.g. famous celebrity deaths), there’s usually a news story on almost any major website. Moreover, such news-ish searches are less important to me than standard queries. My motivation for looking at what’s new is idle curiosity; I’m probably just looking for a way to procrastinate. Traditional searches, however, focus on what I need in order to complete a task I’ve set myself: when I blog, I search for links (like the one above to TechCrunch), and I can’t complete my task till I’ve found what I’m looking for. Real time search is typically not crucial in this way.
Moreover, I don’t understand why it’s such a hot area for startups, since it can’t really make money the same way traditional search engines do. If you think about Google as a business, what they actually do is connect searchers with people who want to sell them things. They only make money when you click on an advertisement. This works well, because oftentimes you look to Google to find information about things you might want to buy (or at least information related to something buyable; looking up asbestosis implies that you should probably sue somebody). Real time searches do not share this characteristic; they pertain to news that happened to someone else, not to you. When you’re looking for information about the Time Square bomb threat or about Tiger Woods’s newest lover, you don’t want to buy Broadway tickets or golf clubs.
This isn’t to say that real time search isn’t useful at all or doesn’t plug a minor hole in search. It’s true (even after the recent real-time ramp-up ) that search engines do badly at finding data based on recent events that aren’t important enough to be newsworthy. For example, I debated in college and have wanted to find out results for the World University Debate Championship (WUDC) currently being held at Koc University in Turkey. Unfortunately, searches for WUDC results, WUDC 2010, or WUDC Koc results all lead to websites that are several months old (despite the existence of a blog reporting the results). Thus, better real-time search would help me a little bit, but not in any game-changing way. Search could, in fact, be measurably improved (a huge amount of search engine optimization (SEO) by websites spammifies results, and I’d love to see a much better product search with constant price comparisons), and it’s a shame that such a minor and feeble trend as real-time search is sucking up so much talent and creative energy that could be better used elsewhere.