TrendSpottr is one of 53 companies chosen by VentureBeat to launch at the DEMO Spring 2011 event taking place this week. These companies do pay a fee to present, but our coverage of them remains objective.
“Real-time search” is one of those buzzword-y phrases that gets tossed around a lot — in part because no one has really nailed the experience yet. Now a startup called TrendSpottr, being unveiled at DEMO today, says it uses real-time search to help users find the hottest content on social networks.
After watching founder Mark Zohar’s demo, it seems like TrendSpottr’s real differentiator is its ability to automatically curate and filter its content. If you’ve ever tried to use the real-time search feature in Google, or to look at trending topics in Twitter, the importance of curation should be obvious — without some kind of filter, the search results are almost useless because there are so many duplicated tweets and random comments.
TrendSpottr, on the other hand, tries to eliminate all the extra noise and hone in on the headlines and other media that are actually driving those tweets and Facebook likes. So when you visit the TrendSpottr site, instead of being confronted with a firehose of content, the results have been arranged to show the top trending headlines, videos, images, and so on.
You can also access TrendSpottr results beyond the website by installing the TrendSpottr bookmarklet. Then, when you highlight a phrase on another web page, the bookmarklet will give you the top trending content related to that phrase.
Zohar added that the company is taking a smart approach to building its infrastructure. Most real-time search companies are trying to bring old-school search indexing techniques to the real-time world by storing as much data as they can. On the other hand, TrendSpottr is focused on the most recent content, because that’s what’s relevant. He said the distinction is comparable to the way that NASA spend “gazillions” of dollars trying to build a pen that works in zero gravity, while the Russians saved money by just using a pencil.
“Both our technology and business model are geared towards massive scalability and distribution,” he said. “Our algorithms are designed to identify and curate relevant real-time content in the most scalable way, without the need for massive amounts of data indexing, server infrastructure or storage.”
TrendSpottr has been in private testing mode, but Zohar hopes to open the site to the public in the next few weeks. He also wants to work with publishers to add TrendSpottr search results to their websites.
The company is self-funded.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.