Owen Thomas nails it:

Whatever happened to Zuckerberg's Law?

In 2008, Mark Zuckerberg told a tech-conference audience that the amount of information people shared seemed to be doubling every year.

It was a bold prediction. It just doesn't seem to be coming true.

If anything, people are undersharing—and that poses a massive challenge to the Facebook CEO's vision of Graph Search.

Zuckerberg unveiled Graph Search, a considerable upgrade to Facebook's notoriously weak search feature, last week, to much fanfare. For the lucky few who get access to it—Facebook is rolling it out slowly—it will be far easier to learn which friends are available for dating, who might be a good recruit, and which restaurants friends frequent.

All that assumes, however, that people actually share that data with Facebook. Graph Search can only search what people share. And it's not clear most users actually share enough to make searches on Facebook interesting..

Facebook has been testing Graph Search internally. But Facebook employees are notoriously heavy users of the service. As early employee Kate Losse chronicled in "The Boy Kings," her memoir of her time at the social network, sharing is expected as part of Facebook's culture.

Consider this impressive number: Facebook users have collectively posted 240 billion photos to date, with its 1 billion users now uploading another 300 million a day. But three years ago, Facebook had 400 million users who were posting about 100 million photos a day. That's about a 10% annualized increase in photo-sharing activity—hardly the doubling Zuckerberg spoke of in 2008.

Location is another form of data that Facebook has tried to get users to share. But that's mostly been a flop. Facebook's standalone Places feature, more or less a Foursquare clone, was killed off in favor of a subtler inclusion of location information in status updates and photo posts. And it turns out that most Internet users—even savvy smartphone owners—aren't interested in checking in.

Facebook has made attempts to get people to provide better local data—for example, encouraging users to rate restaurants after they check in. But that assumes they've checked in at all.

What about work-related searches for recruiting—a category Facebook engineering manager Lars Rasmussen highlighted at the launch of Graph Search?

Pete Kazanjy, the CEO of TalentBin, a recruiting search engine which gathers information about prospective hires across multiple social networks and websites, has actually run the numbers.

TalentBin hasn't analyzed all of Facebook's users. Kazanjy describes its sample set as "heavily biased towards Silicon Valley types who have more of a proclivity to fill this stuff out." Even among that tech set, fully half had no job information at all on their profile.

The other half, on average, provided 120 characters worth of information about their work history. LinkedIn users, by contrast, had 1,800 characters of work information—15 times as much data. 

"There's zero threat to LinkedIn from structured data in Facebook," Kazanjy concludes—because the data isn't there to search on.

So much for finding new hires through Graph Search.

Facebook's most recent attempt to accelerate the pace of sharing, a strategy it called "frictionless sharing," seems to have flopped.

A category of apps called "news readers," which more or less tricked users who clicked on a news story shared by friends to install an app that similarly reported their reading habits, drove users crazy. After specifically recommending the frictionless approach to developers, Facebook backtracked, saying that those features lead to "low quality user experiences."

Outside of limited categories, like music and fitness apps, frictionless sharing seems dead—and with it, the ever-increasing pace of information posting predicted by Zuckerberg's law.

So where does that leave Graph Search?

One theory people have—and until Graph Search is rolled out at scale, it's nothing but a theory—is that once Facebook users figure out that the information they post is easy for friends to find, they'll share more.

But that could just as well go the other way: The ease with which photos, check-ins, and likes can now be found might scare users into sharing less.

Graph Search was a necessary move for Facebook: As Facebook's Rasmussen points out in a blog post, the company had to upgrade the hodge-podge of disparate search engines it had built over time.

But Graph Search is necessary, not sufficient, for Facebook's success. The problem Facebook must solve now is making Zuckerberg's Law more than just a theory.

Oh and by the way, people still freak out when someone publicly shares their private Facebook statuses, as Storify just found out the hard way:

http://www.businessinsider.com/storify-pulls-in-private-status-updates-from-facebook-too-2013-1

Whoops. Sorry, Zuck, consumers are resisting some of your theories, bro.

I think he has helped in educating how open the internet is and making people more distrustful (a good thing)...there is that hard objective reality where governments track all the ones and zeroes.  

Yes, he has given people a healthy skepticism of the notion of privacy in the Internet age.

Graph Search + labyrinthine privacy settings = complete disaster.

Zuck has always known that in order for Facebook to work, a large percentage of its users would have to be heavy sharers. So FB has consistently pushed people to share more and more data, even as it has had to add more and more privacy settings to make users feel safe about uploading personal information.

And since that social data is less valuable when locked down with strict privacy settings, FB has had to periodically violate users' trust and make all that information more public.

The usefulness of the new graph search feature will be limited both by lack of sharing and lack of public sharing. I believe that they're going to have to pull another of their famous sleazy maneuvers and violate users' trust once more.

Graph Search probably sounds great in a board meeting, and no doubt demos well on FB employee data, but I thinks it's just not going to work at scale. Users don't want to share enough of the necessary data to make for good results that are also restricted by social graph. And to deliver good search results, FB will have to start making searches more public. They may start by using "anonymized" aggregate data, but they'll eventually have to struggle with the balancing act they've been playing between user trust and making the site useful.

In a lot of ways, FB is an unnatural product. Zuck's interests have never been aligned with those of users, and everyone has suffered as a result.

Now as always: on Facebook, you are the product.

If he could somehow abandon advertising as a business model in favor of commerce, payments, and/or value-added subscription services, everyone would be better off.

oh no...so I am the product here on PW too?

Heck no. We don't have revenues on PandaWhale. :)

And one day, when we do have revenues, they will not be based on advertising.

They will be based on commerce, payments, and/or value-added subscription services.

I love freemium products like Dropbox and Evernote. We'll come up with something like that.