Follow the Filter Bubble

The future of personalization is here


/ June 7th, 2011 /

Writing The Filter Bubble in 2010, Eli didn’t lack current examples of how personalization algorithms shape the content we see on our browsers – from Netflix and Amazon suggestions to Google search returns and ads that trail us around the internet.

But many of the creepiest applications of the filter bubble were left to the imagination. What if, someday, search engine results changed depending upon your friends’ preferences? What if you visited your favorite online magazine and the front page was “personalized” to show you the stories it thought you wanted to see?

Well, it didn’t take long for “someday” to arrive.

Last month Bing started giving users the option to plug into to their Facebook accounts to “receive personalized search results based on the opinions of your friends,” thus ”bringing the collective IQ of the Web together with the opinions of the people you trust most, to bring the “Friend Effect” to search.” Such a move was expected when Bing and Facebook joined forces last year, but it’s still something to see it in action. To Bing’s credit, the “Facebooking” of search is opt-in (at least for now), so searchers can decide for themselves how much they want their Filter Bubble to collapse in on them.

It remains to be seen whether Washington Post and Slate readers will be given a similar option. Trove, a project of the Washington Post, creates a personalized news experience based on your Facebook profile. It’s still a separate site and in its early stages, but from an interview with Chief Digital Officer, Vijay Ravindran, it looks like the Washington Post will be integrating parts of Trove into its own site and subsidiary Slate in the not-too-distant future. It shouldn’t be long before my Washingtonpost.com is something quite different from your Washingtonpost.com.


Twitter Search Results Now Personalized


/ June 1st, 2011 /

Twitter — a last outpost of the unpersonalized internet — has taken the plunge.

Twitter search results, previously sorted solely based on time, will now be personalized.

Today we’re starting to roll out a completely new version of Twitter search. Not only will it deliver more relevant Tweets when you search for something or click on a trending topic, but it will also show you related photos and videos, right there on the results page. It’s never been easier to get a sense of what’s happening right now, wherever your curiosity takes you.

There is a bit of irony to the way twitter is marketing this announcement.  Their blog boasts that the new search features can take you “#anywhere.”   They’re pitching the change as enabling discovery in spite of the tendency of personalization to do the opposite.

In the end, the devil is in the details.  How are results being personalized?  Is way in which results are being personalized clear to users?  And most importantly, how much control do users have over personalization?

Danny Sullivan has a quick rundown of Twitter’s approach here.

How does Twitter decide what’s most relevant, what should show up as a Top/Hot Tweet?

“Relevance for us today is using a combination of signals, your follower graph, who you follow, who’s following you. Another aspect is just looking at the content itself and the resonance of the content,” Mike Abbott, Twitter’s vice president of engineering who oversees search, told me.

So Top/Hot Tweets are personalized for each individual? Yep. But those worried that this means you’ll only see things from those you know or like, relax. The personalization isn’t that dramatic, Abbott says.

Indeed, Abbott showed me a screenshot of his own results for a search on a topic about a Twitter feature that showed some tweets that weren’t in favor of it. It’s not just reinforcing what you like.

If you don’t want personalized results, you can log out of Twitter and search that way. Then you’ll get “normal” results, the company says.

At the end of the day, this seems to be a relatively responsible approach to personalization that improves user experiences while also preserving users ability to get unpersonalized results.  Unlike Google, which personalizes results whether you are logged in or not, Twitter has preserved the option to log out and get “objective” or “normal” results.


How to go bubble-hopping


/ June 1st, 2011 /

Perhaps the most unnerving thing about the filter bubble is its invisibility. Even if you’re aware that online filters are distorting your perspective, it’s almost impossible to know what the Internet would look like if you weren’t, well, you.

But a new Twitter feature is making it easier to see things from a different vantage point. Twitter users can now view not only the roster of people another user is following, but also the combined stream of those folks’ tweets. In other words, one user can now see Twitter through the eyes of another.

And Twitter’s not the only service offering this kind of bubble-tourism. Marshall Kirkpatrick at Read Write Web points out that News.me, the New York Times‘s and Bit.ly’s iPad app, lets you view the article feeds of other users (provided they’ve opted in). “The app makes it remarkably easy,” Kirkpatrick writes, “to find yourself reading content published and shared by people with world-views different than your own.”

Features like these could serve as an antidote to an increasingly segmented online universe; hopefully we’ll start to see more of them. If each of us is stuck in a personalized bubble online, it’s at least nice to be able to visit each other.


Love in the filter bubble?


/ June 1st, 2011 /

Personalization algorithms already tell us what movies to watch, news stories to read and tunes to listen to. It was only a matter of time, then, that they’d tell us who to love.

Matching algorithms aren’t new to online dating services. EHarmony, Chemistry and OKCupid have long served up compatible mates based on dozens, if not hundreds, of questions singles answer on their sites.

But a new dating app, StreetSpark, is venturing out internet-wide to pick up clues on who you’re likely to become enamored with. Love seekers on the site can plug into their Facebook, Foursquare and Twitter accounts to discover potential lovers with similar tweets, profiles and cafe haunts. (That, at least, is the concept. So far this single has yet to be sent a match.)

It’s like “traditional” online personalization but in reverse. Instead of telling you what you’ll like based upon your friends’ preferences, it tells you who you’ll want to be friends with based on what you like.

StreetSpark touts their service as giving “serendipity a helping hand.” Normally we have to wait for luck to bring us face to face with that special someone; StreetSpark provides us with a helpful homing device right in our smartphone.

It’s an odd usage of “serendipity,” though, which describes the phenomenon of making desirable discoveries by accident. If you instruct your iPhone to tell you when there’s a sympatico mate in your hood, bumping into them can’t really be described as “an accident.” Of course, the makers of StreetSpark are aware of that contradiction and are tongue and cheek in using the term.

But it’s more than a semantic quibble. Part of appreciating the beauty of “making discoveries by accident” is to understand that sometimes we don’t know what we’re looking for. If you’re a romantic, that can especially be true in the case of love. It’s not as if we have the profile of “the perfect guy” in our head and falling in love is just a matter of luck when you’ll run into that profile. The “accident” of love is when we meet someone who doesn’t fit our pre-conceived ideal and yet, mysteriously, we fall head over heals. In the process, if we’re truly lucky, we’re opened up to a new, exciting and unknown world.


Select your 150 friends wisely


/ May 30th, 2011 /

Online technol0gy is transformative. It can make the world flat, spark revolutions and even wrap us into personalized filter bubbles. But there’s one thing technology hasn’t been able to do yet: expand our circle of friends.

You’re probably familiar with Dunbar’s “150 rule:” the reason that humans tend to limit the size of their communities to 150 people – whether in prehistoric towns, in military units or in cults – is because the human brain maxes out at 15o friends.

Now it turns out that even Twitter can’t free us of this 150 ceiling. Bruno Concalves and colleagues at Indiana U recently looked at 1.7 million tweeters over 6 months to see how many connections they kept up (connections, as opposed to mere followers, had actual back-and-forth exchanges). True to Dunbar’s prediction, twits generally don’t maintain more than 100-200 friends.

That’s bad news for the filter bubble. You can imagine one hope of avoiding a personalized information bubble is to widen your circle of friends in order to include folks with different viewpoints. That way you might expand the information that arrives on your laptop screen. But, as Eli points out in the intro to his book, merely adding friends to your FaceBook list doesn’t mean you’ll interact with them. No interaction means those “friends” will be virtually invisible on your feed. You’re still stuck in a community of 150.

The only way to truly escape the bubble may be to replace some of your current connections with people who disagree with you – and actually engage them in discussion. Of course, if we’re concerned about the limitations of human behavior, that may be the most pollyanish hope of all.


Subscribe

Events


Tweets


    Posts


    Categories

    New York Times Bestseller

    Buy the book from

    Amazon.com (US)
    Amazon.co.uk (UK)
    Barnes & Noble
    Indiebound

    Buy the audiobook from

    Tantor Audiobook

    "Eli Pariser is worried. He cares deeply about our common social sphere and sees it in jeopardy. His thorough investigation of Internet trends got me worried, too. He even taught me things about Facebook. It's a must-read."

    David Kirkpatrick, author of The Facebook Effect

    View More

    You can reach Eli by email at epariser@elipariser.com