Exciting news these days from SemantiNet. Over the past year, the company has been developing an application for online publishers based on its semantic web technology that will allow publishers to automatically enrich the content on their sites.
SemantiNet’s Headup for Sites has been running on several hundred blogs for the last six months. Now, SemantiNet has launched its first installation with a major publisher. They completed a successful implementation at the Jerusalem Post, which is currently running a wide-scale pilot.
So, how does it work?
Headup scans Web sites and builds up a database of significant entities (people, places, and things). It then automatically generates Topic Pages which combine articles from the site and enriches them with additional content from external sites such as Wikipedia. Finally, it goes through and annotates instances of those entities on the publisher page.
For example, here is an article from Jpost.com which has been annotated:
The user who hovers over the link (in this case, “Hosni Mubarak”) sees a popup snippet that provides some information about the entity and a link to its topic page. Clicking on the link brings up the Topic Page:
Why is this important?
Online publishers, especially in the news world, face a major engagement problem. Studies have shown that users are keen to consume news online. In fact, it is a more popular online activity than shopping, watching videos, or social networking. However, user engagement on news sites is low and the bounce rate is high.
By utilizing Headup, news sites provide their readers with a more complete context as well as an easy way to consume additional content. By creating topic pages, publishers can essentially unearth archived material that users would generally not think to look for and present it to their users within a particular and more complete context. Users get more information about the articles they are reading and sites get more pageviews and provide a better user experience.
This is the reason that the biggest online news publishers such as the New York Times have also developed systems for creating topic-based directories of archived content.
The cool thing with Headup is that it allows publishers to develop these topic pages automatically. In addition, its semantic technology identifies and categorizes entities based on their context rather than on keywords. For example, it knows that stories mentioning “Benjamin Netanyahu”, “Bibi”, and “Prime Minister of Israel” all refer to the same thing and can organize topic pages containing all three.
In addition, the technology recognizes related linkages between entities and can then present the user with even more content around a general subject. For example, the same topic page on Netanyahu could lead to material on Knesset Members or Ehud Barak or even Sayeret Matkal.
SemantiNet believes that the future lies in bringing new levels of meaning to the content in the online world. Headup is a significant step in that direction.