Posted by Jeff as Uncategorized
We did two things this week that are interesting for those of you interested in how behavioral data is being used to drive new feature development in RSS.
First and foremost, we released Inbox 3.0 for Microsoft Outlook and it takes advantage of your attention stream in the form of APML to organize your feeds based on what you give the most attention to, in other words your reading habits.
By combining this “relevance metadata,” NewsGator can provide their users, like those using Inbox 3.0, with the most relevant news and information while also keeping track of what subscriptions you care most about and allowing you to export this information in APML (Attention Profiling Mark-up Language) so it can be shared with other services.
Shortly after releasing our free consumer clients I wrote a post about attention data that explained in some detail the distinction between attention streams focused on people vs. content. The above is an example of how we are using fully portable attention data streams in the form of APML to deliver new functions to users, but are not locking our customers in to our products. The attention streams around content are equally interesting to us.
Why is this attention data useful? Simply put, attention infers content authority and quality; if you share something I can make an assumption that you found it useful, which we can then use in our attention algorithm. The scoring generated by our attention algorithm can be used to make search more accurate, and it can be packaged as an API that we make available to our partners to enable their services to better filter and sort content.
To the point that we have attention streams that are content focused, we have been busy developing new features that expose this. We quietly turned on Activity Scoring in Feb and have been improving it since then. Simply put, we started feeding all those clicks, tags, bookmarks, and clippings events into an algorithm that scores each and every post in our network, which has grown to over 8 million new content items daily.
In NewsGator Online there is a sorting option that will sort the feed based on this Activity score. My personal experience is that it’s mildly useful but we’re not going far enough with it. It’s in the aggregate that activity scores become truly useful, such as sorting my clippings folder or looking at a group of feeds from an activity perspective.
The goal of activity scoring has never been to develop the be-all-end-all algorithm. We have always looked at this starting point as just that, a starting point, and it has further been our intention to both make this scoring available to others as well as integrate other RSS attention scoring measures with our own.
To this end we announced today that we have enabled another attention sorting option based on AideRSS. First a little background, Jevon MacDonald introduced me to Ilya Grigorik last July when they were first getting off the ground. We got more serious about doing something together when I joined NewsGator (I literally called Kevin and Ilya my first week with the company). I love what they are doing and use it regularly myself, but it’s important to recognize that what the are doing with Postrank is fundamentally different than what we are doing with Activity Scoring. We build a score from within our network, they build a score from external factors, so the two approaches are very complimentary.
Right now we are offering AideRSS as a sorting option but this is just the start. We are also working with additional companies to bring integration of attention streams to bear on the broader problems of relevancy and discovery in RSS. Longer term I think there is a new generation of RSS products that we, as users, will benefit from in that they will go beyond the basic workflow of subscribing to a feed and then consuming a chronological stream of posts. There remains a great opportunity to present a techmeme-like interface that organizes content based on the content subject, sentiment, and related content. Lastly, we’re also bringing these features to our widget services so stay tuned!