outbrain

Serendipity feed update

Those of you who subscribe to our serendipity feed (the non-personalized feed), may have noticed some questionable (or - *very* questionable...) posts lately. The reason for this - a surge in new users via our partner API - is a good reason. But the resulting posts are not.

We've therefore decided to freeze the serendipity feed until we can improve the algorithms to properly handle this surge of users. You don't need to do anything on your side... you simply won't see any new posts coming in during the next few days.

Stay tuned for updates on this.

Serendipity feed

Today we're releasing a new feed - the outbrain serendipity feed. You can subscribe to it here:

http://feeds.feedburner.com/outbrainSerendipity

The serendipity feed brings you posts that are starting to gain some steam within the outbrain community, but do not yet have enough rating data to be recommended with reasonable confidence.

What's this feed good for? A bunch of stuff:

1. It's a great way to see newer posts as they start picking up positive ratings in our community.
2. It's a great way to get content from a wider variety of sources. In your personalized recommendation feed we always serve those items with the highest personalized scores, even if they ALL happen to be from Seth's blog... On the serendipity feed we enforce more variety.
3. It's a great way to help other outbrainers get highly relevant recommendations. The more ratings we have on specific posts, the more intelligent recommendations we can deliver. The serendipity feed helps us get those deeper clusters of ratings on specific items.

So go ahead and point your feed reader to the serendipity feed, and start rating those items to make your outbrain smarter!

(Thanks Svec for the idea!)

New algorithm - better personalization, more serendipity

The new recommendation algorithm we've been working on is now live for all outbrain users[1]. With the new algorithm we tried to touch one main point - better personalization. Our previous algorithm would find similarities between outbrainers based on a small number of similar positive votes. Once two users were associated by votes, it was difficult for those two users to drift apart. It's like two random people in a theater who happened to like the movie, finding out as they're exiting the theater that they are now married on that merit...

The most noticeable effect of this 'happened-to-like-the-same-movie' syndrome, was that recommendations became very monotonous and limited to those few blogs that your newlywed spouse happened to like... ;-)

The new algorithm is much more dynamic and sensitive to people's similarities. This means that every time you vote, whether high or low, outbrain will adjust itself to find those people with the most similar interests to you.

Beyond the much better personalization, you should notice much more serendipity in the recommended items. This is something that should further improve with each new user we add into the system.

We'd love to get your feedback on the new algorithm. The best possible feedback is simply by voting on your recommended items every day. We look closely at those votes and tweak the algorithm settings constantly (+of course, the more you vote, the smarter your personal outbrain will get!).

Any other feedback? Shoot us an email to:
info (at) outbrain (dot) com

[1] If you haven't subscribed to your personal recommendation feed yet, do so by pointing your RSS reader to the following location (don't forget to change the USERNAME part with your outbrain user name):

http://dsf.outbrain.com/dsf/USERNAME.xml

New recommendation algorithm coming

We've been quiet, but working hard on a new recommendation algorithm that focuses primarily on improved personalization. We'll be rolling it out within ~2 weeks. If you want to get some results ahead of the release and help us tweak the algorithms - drop me a note at:
galai (at) outbrain (dot) com

Full body posts

You may have noticed in the past few days that your recommendation feed now sometimes contains the full posts (rather than just the titles). This is still a work in progress, but the idea is to serve you with the entire post in your RSS reader instead of having you need to click through to the blog to read it.

If you haven't subscribed to your personal recommendation feed, you can do so by adding this feed to your RSS reader: (make sure to replace the [USERNAME] part with your actual outbrain user name... ;-)

http://dsf.outbrain.com/dsf/[USERNAME].xml

(if you're having trouble subscribing to your recommendation feed, drop us a note at support at outbrain dot com

What's up with the relevancy?

Some of you may have noticed some questionable relevancy in the recommendations feed lately (mostly around Flickr photos, and posts from my blog - Web X.0).

This is caused by 2 factors:
1) We're constantly messing around with the algorithm.
2) We opened this up for very few users until we're more comfortable with the product and the algorithm.

You should notice a significant improvement on this in the next week or two, as we send invites to many more users who have signed up for this. Also - In ~2-3 weeks we plan on rolling out a brand new shiny algorithm that should greatly improve overall relevancy and personalization quality.

Thanks for your patience as we try to make this a great product, and... - keep voting!   

Personalized recommendation feeds

Outbrain is all about 2 things: floating the best articles/posts and flagging the worst. And most importantly - personalizing the recommendations and filters to each user's personal interests.

This week we're releasing outbrain's first recommendation feature - a personalized recommendation feed. After subscribing to your feed (details and links in an email later this week), we will drop up to ~3-5 posts-per-day which the outbrain algorithm found to be potentially highly interesting for you.

We do this by looking at historic votes, and finding people with similar voting patterns. Therefore, the more you vote on items you read, the more intelligent outbrain will get about the recommendations it gives you.

This feature focuses on recommendation - finding those nuggets that you would otherwise probably miss if you had to read all those posts out there all by yourself. Our next feature will focus on filtering - flagging those items that the collective outbrain intelligence thought were a waste of your time. Stay tuned.