Pop Your Internet 'Filter' Bubble

Think you're on the Internet right now? Well, you're not. You're on your Internet. The exact version of the online world that you see is available only to you.

Most of the major conduits through which you see the world online, including Google Search and Facebook, are gathering all kinds of data about you, then presenting you with a custom version of the world they think you're interested in. They're hiding or de-emphasizing the version of the world they assume you're not interested in.

In the past two years, the biggest gatekeeper websites have gotten very good at figuring out what you want and giving it to you. What's wrong with that?

There are downsides, according to a new book called The Filter Bubble: What the Internet Is Hiding From You.

In a nutshell, the book argues that the sophisticated personalization engines improve things in the short run. But in the long run, they dumb us down, reduce our exposure to new ideas and ultimately could lead to a society without a shared set of facts about the world. The personalized Internet favors the marketers and propagandists but provides an obstacle for people who are trying to introduce new ideas.

The Internet is increasingly turning us all into dictators and divas. Like the entourages of Saddam Hussein or Jennifer Lopez, the Internet tells us what we want to hear, feeding us a version of the world that feels good, goes down easy and doesn't challenge us.

The book ships May 16. It was written by Eli Pariser, who is the president of the MoveOn.org board. MoveOn is a liberal public-policy group, and Pariser's concerns are mainly political. But the "filter bubble" concept affects you no matter what your interests. And you're going to hear a lot about this concept after the book hits.

In this column, I'm going to tell you how personalization works, why you may not want it, and also how to pop the bubble and opt out of a system that censors your Internet based on stereotyping.

Your own private Google

The "secret sauce" of Google Search has long been an algorithm called PageRank (named after co-founder Larry Page). But on Dec. 4, 2009, Google announced an additional algorithm that custom-tailors search results according to the individual attributes of the user.

According to Pariser, Google uses 57 "signals" -- even when you're not logged in to Google -- to customize search results. (Google was unable to confirm the number of signals.)

These "signals" include where you are, what you have clicked on in the past and who your friends are. But that's just the beginning. Google also gathers information about which browser and device type you use, how much you travel (based on where you use search over time), how long it takes you to click after getting a search result, and many, many other data points.

From all this data, Google decides how to sort your search results. (A Google spokesman told me the company rejects the term "filter" because it implies that Google is hiding links rather than prioritizing them.)

Here's a fun experiment to try. Search for something on Google, and have a friend or two do the same search. See how the results are different? Many of the links are the same. But they're in a different order and "skewed" subtly in one direction or another.

As we increasingly use mobile devices, our exact locations, where we've been, which stores and restaurants we've entered, who we've met with and possibly even activities on unrelated services will be increasingly factored into the decision-making process about what we see and don't see. In fact, all the major improvements Google has promised for search involve far more and deeper personalization.

"The power of individual targeting -- the technology will be so good, it will be very hard for people to watch or consume something that has not in some sense been tailored for them," according to Eric Schmidt, Google's executive chairman.

Google apparently tries to be responsible with its power over your attention. A spokesman told me Google understands that "people value diversity on results pages." The company uses its deep understanding of you not only to target your assumed interests, but also to deliberately challenge you with a few links outside those interests.

Facebook's antisocial filter

Facebook is less responsible, in my opinion. The social networking giant determines what appears in your "News Feed" using an algorithm called EdgeRank. (Facebook ignored my request for an interview.)

Every action you take on Facebook -- clicking "Like," commenting, sharing, etc. -- is called an "Edge" internally at Facebook. Each Edge is weighted differently according to secret criteria.

What you need to know is that relationships and content that don't get enough "Edges" will get "edged" out of existence. Facebook will cut your ties to people -- actually end the relationships you think you have -- and block content that doesn't earn enough Edge points.

For example, many Facebook friendships exist solely through reading each other's Status Updates. An old friend or co-worker talks about a new job, shares a personal triumph like reaching a weight-loss goal, and tells a story on Mother's Day about how great his mom is. He posts and you read. You feel connected to his life.

Without telling you, Facebook will probably cut that connection. Using unpublished criteria, Facebook may decide you don't care about the person and silently stop delivering your friend's posts. Your friend will assume you're still reading his updates. You'll assume he's stopped posting.

Any friends who fail to click or comment on your posts will stop getting your status updates, too. If you have 500 friends, your posts may be actually delivered to only 100 of them. There's no way for you to know who sees them and who doesn't.

Facebook also filters content. EdgeRank keeps track of how many of your friends comment on a link to content, and it will use that criteria in the default view of your News Feed, which is the "Top Stories" setting.

The vast majority of even technical, savvy people I asked about this have no idea that their friends' activities are determining what content they see and don't see on Facebook.

Why everyone is doing it

It's not just Google and Facebook that shape and filter what you see online based on invisible assumptions and behind-the-scenes stereotyping. Amazon, Netflix, Pandora and hundreds of other companies offer "recommendations" or content based on personalization algorithms.

And personalization is becoming big business. Companies like Strands license their personalization engines to other companies. Strands customers include major banks, coupon and discount services and retailers, music sites and advertising companies.

The whole Internet is rapidly being personalized. Nobody can predict what kind of Internet -- what kind of world -- will emerge when everyone has a unique view of the world that nobody else can share.

Companies are aggressively pursuing personalization because it makes users happy. Personalization validates existing beliefs and prejudices. "Consuming information that conforms to our ideas of the world is easy and pleasurable," according to Pariser. "Consuming information that challenges us to think in new ways or question our assumptions is frustrating and difficult."

Personalization can create an "identity loop," according to Pariser. If you click to satisfy some passing curiosity, the algorithm might favor more such links in future. Because there are more links, you click more. You might even monitor your own activity and conclude that you must be especially interested. Personalization not only responds to personal interests. It shapes them.

And personalization based on activity favors the frivolous and the commercial. We all click mindlessly for temporary escapism. But we don't realize that we're training the Internet to favor that kind of content over important information.

Ultimately, personalization is ideal for marketing. We want perfect relevancy when shopping. As one venture capitalist said at this week's Social-Loco conference in San Francisco, "when you walk into a store, the only shoes and clothes available should be in your size."

The Googles and the Facebooks of the world are advertising companies. Their customers are advertisers, not users. And their customers love user personalization, because it's the shortest line between consumer and point of sale.

Of course, most sources of content are "biased." The site you're reading now, for example, is "biased" in favor of technology-related content over, say, stories about Latin music. The difference is that online personalization is invisible. Nobody knows what's being filtered out or why. Most people don't even know that filtering is taking place.

How to pop the bubble

If you don't want your Internet filtered by some invisible stereotype, here's how to pop the bubble. These tips are a combination of my own, plus some offered in The Filter Bubble.

* Deliberately click on links that make it hard for the personalization engines to pigeonhole you. Make yourself difficult to stereotype.

* Erase your browser history and cookies from time to time.

* Use an "incognito" window for exploring content you don't want too much of later.

* Use Twitter instead of Facebook for news. (Twitter doesn't personalize.)

* Unblock the Status Updates of your friends that Facebook has already blocked. Click the "Edit Options" link at the bottom of your Facebook News Feed. The dialog box will show you who is being blocked. You can hide or un-hide each friend manually, or unblock everybody. This dialog box affects only what comes from friends to you. It does not affect what your friends see of your posts.

* Every week or so, post something and then ask the Facebook friends you really care about to go "Like," comment and click. This activity should prevent Facebook from censoring your comments later for these people.

The most important thing about the "filter bubble" is that you know it exists. The Internet you see is not the Internet I see. The Internet you see has recently been redesigned to flatter, pander and validate -- not challenge, enlighten and educate.

The filter bubble is real. But it can be popped.

Mike Elgan writes about technology and tech culture. Contact and learn more about Mike at Elgan.com.

Subscribe to the Daily Downloads Newsletter

Comments