Real-Time Search Is Dead, Sort Of

So, it looks like real-time search has died. The two leaders in this space, OneRiot and Collecta, have decided to focus on other things. OneRiot announced their change in focus last fall as they launched their advertising network. Now we hear that Collecta is closing their search as they change their focus as well. In addition to the announcement, there is an amazing quote in that post:

Asked whether creating a real-time search engine is a viable start-up business, Campbell answered quickly: “No.”

So, the CEO of one of these companies decides to “put the nail in the coffin” for real-time search. The real question is why is real-time search dying? My thinking is that real-time search, positioned as Google for real-time streams, is a solution looking for a problem. Ask yourself, do mainstream users care about real-time search? First, they would ask you what real-time search is. Then when you told them you would probably get the ever-popular “deer in the headlights” look. If you ask early adopters, they may say the future is all about real-time search. Early adopters are somewhat correct, but not quite.

There are a number of issues restricting adoption of real-time search. First, what problem does it really solve? If we ignore mainstream users, we may have the problem of know what is news now, and when did it happen. However, there are very few people in the world that really need to know that. Also, people who want to know about things as they happen already have Tweetdeck, or their favorite real-time client, on a second monitor or at least running constantly. So, they probably will know that something has happened shortly after it has occurred.

There is an additional problem that real-time search has difficulty getting traction because there are not a lot of real-time data sources. In reality, there is a small handful like Twitter and Facebook, and most people that care will watch their stream throughout the day. If you watch your stream already, why would you go to another site to execute a search? This is the same reason why Google works so well, there are millions of sources of web pages for any topic you can think of. The aggregation into a search engine is required in order to discover useful pages. Real-time Google-like searching just does not make sense.

Another reason for the lack of success is that if something is somewhat important to a group of people, the item will be repeated several times throughout your real-time stream. This repetition will ensure that you notice the item if you are watching the stream. This also points to why real-time data is important. Trends. This is where real-time data shows its worth. Trending topics are a marketing gold-mine. Trending topics shows you what people are talking about or at least what they have been talking about for the past hour.

Trending topics also suck right now because they are being gamed mercilessly. Have you looked at any of the trending topics on Twitter? Invariably, by the time something becomes a trending topic there have been plenty of spammers using that topic to spread porn links and other spammy things. However, if we ignore the specific items that people are posting and take a Google Insights look at a trending topic, like the one for a current trending topic of “catwoman”, then you have some interesting information. You can see what surges of interest occurred and what news items they may have been related to.

That type of trend information is powerful and interesting. When you get to that type of analysis, you can compare various trends, whether it is current year vs. previous year or your product vs. a competitor’s. Some of the social media monitoring applications allow for this type of analysis, but you are paying a significant amount of money for it. Also, you are likely limited to select searches as opposed to the general purpose and open search of an engine like Google. So, why isn’t there a general purpose engine like this? Because there has not been a need, and trend analysis is an expensive function. We do have some special purpose engines like Topsy, that are focused on a specific site like Twitter or Facebook. However, these engines do not take into account the social aspects of these streams.

Google is trying to make search more social by including relevant items from your social contacts. They are still struggling with getting good results. Facebook is trying to own social search, but it has not been their entire focus yet as they continue to struggle with privacy concerns and fine tuning their social stream. Twitter has practically ignored their search engine, as it has not changed much since its original purchase. Maybe Topsy will evolve into this type of solution or maybe another player is being built. In any case, the future of real-time search is coming, with all of its social-enabled goodness. Now we just need to wait for it.

Enhanced by Zemanta

5 thoughts on “Real-Time Search Is Dead, Sort Of

  1. Hi Rob,

    You’re right there is someone else coming onto the scene with real time curation of content! If you remember, the site that brought you the TweetMeme retweet button, then you probably won’t be surprised to hear that there is a spin off brand from that called DataSift. DataSift is a real time content curation engine focusing on social web content such as Twitter, WordPress, SixApart, Feeds, Google Buzz and so much more! It also has an augmentation layer bringing together additional services such as Klout, PeerIndex and Salience for identifying sentiment, influnce and reach.

    The service is currently in alpha and you can sign up to the alpha programme at the moment.

    Keep an eye on our blog for regular insights into what we are up to and to see the features of our service in action.


Comments are closed.