As you might have guessed, I love trying out new sites that pop up on the Internets. The vast majority I check out once, and never return. However, the one place that I’ve been going to more and more over the past few weeks has been the new Twitter search functionality that was built into Bing.
Full disclosure for those who don’t know – I do work at Microsoft. But putting silly religious search engine wars aside, it is a cool service and will likely be implemented in a similar way when it rolls out to the search engine of your choice in the near future.
Anyway, it’s a very useful service and got me thinking about real-time search. Before checking out Bing’s Twitter search, I had assumed the problem that needed to be solved was something like this: now that we now have a ton of real-time content (tweets), how do we index, weigh importance, and filter the content to discover the most interesting tweets?
However, after seeing how much more useful Bing’s Twitter search is compared to Twitter’s built-in search, I started to look at the problem differently. The key is that the tweets themselves aren’t the content, they are simply indicators that can help determine what links to non-Twitter content are most important for a given topic, at a given time, in a given location.
If you look at it from that perspective, the problem is now reversed. Instead of trying to build a real-time search engine, what about creating a real-time content engine? That’s the basis for today’s idea, Qwiki.
Here’s how it would work. An algorithm would monitor Twitter to determine the newly-trending topics. Once a topic hits a certain threshold, the Qwiki site generates a web page for that topic with as much information as it can automatically pull from both Twitter and the web.
Qwiki would then tweet out a link about that topic. As people come to the Qwiki site, they could start to fill out information about the topic, and retweet the link out to others. As more and more people come to the page and contribute information in real time, the wisdom of the crowds will take over and page will quickly become the best source of information about that topic.
As a result, just as Wikipedia is one of the top entries on most search engine queries, Qwiki results could rise to the top of real-time search requests. And while I usually don’t like ad-based revenue models, it might actually make sense for this one – especially if location-based, real-time advertising begins to take hold.
Speaking of real-time, I’m in Mountain View today and will be working out of Red Rock Cafe from about noon to 2:30. If you’re in the area, swing by and I’d be happy to buy you a coffee. I’ll be on Twitter, you can contact me there at @astartupaday.