Here's a post I just wrote for Techcrunch. Pick your location to read & comment :)
Google is an algorithm driven-company. "PageRank" (named after +Larry Page
himself) was the "founding algorithm" of Google -- the one that gave it superior search results, and eventually led to Google "winning" the search wars of the early 2000s. The algorithm continues to evolve -- in fact, it's Google's most important work -- and by some accounts, it includes more than 500 million variables and 2 billion terms to perform its magic. Can a company so enamored with the power of algorithms and machine learning, let the user take control? This might be a more precise way of raising the question people keep asking. Is "social" in Google's DNA?
I love using G+, enough so that I'm worried that Google is going to make a misstep and ruin the service. Specifically I worry that Google will assume an algorithm alone is what's needed to reduce the "signal to noise" ratio in the G+ feed. Several Google engineers have posted publicly that they're working on this algorithm. I've been making my opinions known in comments for a few weeks now -- hoping to catch the ear of Google engineers, but now that it's harder to gain their attention as 1 voice within 10 million, I thought I'd do better to post something more substantial.
One of the key issues that will determine the fate & nature of G+ is whether Google favors an algorithmic approach over a user-controlled approach to the stream. Facebook (almost counterintuitively) is the one that favors an algorithmic approach, and currently it’s one of the defining differences between Facebook and G+. As usual, +Mike Elgan
nails this one. As Elgan wrote recently: "Facebook deals with information overload by using a secret algorithm to judge the quality of your relationships, then secretly blocking most of your updates to your friends" And: "Google+ deals with information overload by giving you, the user, real control. " At least that's what G+ does right now. Is it going to stay that way? And behind the scenes right now, are G+ engineers working more on their sorting algorithm or on features that would further enhance the users' control of the feed?
Because of the way my posts, articles and my very presence on Google+ have been interpreted, people seem surprised when they learn I love Facebook as much as G+. I'm rooting for Zuckerberg & Co. as much as I am for Google. I want to see more and distinct networks thrive. I don't think social networking is a zero sum game. I suspect that people believe that social networking is a "winner take all" endeavor, because they mistakenly assume people "left MySpace for Facebook." Facebook didn't kill Myspace; MySpace "committed suicide" through continual mismanagement. (For what it's worth, I include myself in that group of mismanagers :-) and I don't mean to blame any single individual -- the troubles were tremendous. I'll explain it somewhere else, someday.) Likewise, MySpace did not "kill Friendster" -- Friendster had its own set of problems. If they'd been corrected, I believe both MySpace and Facebook would have thrived as different types
of social networks. (In fact, Friendster basically would have been "Facebook" -- a real name network, focused on real-world relationships for efficient communication.)
Anyway, I love using G+ and
Facebook. On Facebook, nearly all of my "distant" friends and former co-workers are there. It's the best way to keep in touch with them. But recently I've noticed that I get less and less response from my Facebook friends. I post something that used to generate some interaction, and now I receive almost nothing. I suspect that this has to do with the way the Facebook feed works. And I've done a few tests that seem to confirm my belief. For my own Facebook "consumption," I choose the "Most Recent" Feed option. (For those unaware, it's the way to see everything being posted to Facebook as opposed to what Facebook thinks you want to see.) I've also created some lists and filter my feed to see what certain groups of people are posting. (Yes, Facebook has friends lists, and yes you can share to that list and choose to see only what that list posts. Sounds like "Circles" doesn't it?) The problem is that almost no one else on Facebook does this. And that's why Facebook created groups and uses the "Top Posts" algorithm (the technical name for it is EdgeRank: http://tcrn.ch/npaXax
). Facebook has tried to find different ways to bring the information users want to them, because the "Circles" concept when implemented at Facebook in 2008 ("Lists") didn't work. Facebook has assumed that users can't handle the overload of information, and that EdgeRank is better than the "Most Recent" option. They've downplayed Lists, sorting and Newsfeed "preference" options more & more, so that most users don't even know they exist.
Is that the right move? I'm not so certain. And I'm wondering what kind of data Facebook has to suggest that it is. Mike Elgan covers the "dangers" of EdgeRank in his provocatively titled article "How Facebook Secretly Ends Your Relationships" (http://bit.ly/pIIIpl
) And in it, he concludes that transparency and education is key to helping users understand what is happening in their newsfeeds. Facebook may have already made the decision on which way they want to go, but the G+ team would do well at this juncture to consider the specific suggestions Elgan makes. And even more important, Elgan raises a point he made years ago from another article: http://bit.ly/n4qRHa
-- that the real utility of social networks is not to help us connect with the 10 people in our lives who we care and love about the most, but rather the value lies in being able to cast our net wide & far so we can maintain relationships with 100s of people in a way we never could before this technology. Social networks may be more valuable to us in that they allow us to maintain more "weak ties" than we ever could before. Our "strong ties," after all, are already "strong," and don't benefit as much from the technology boost. If you question the value "weak ties," how about job networking, dating (a shocking (to some) percentage of of new relationships start online), advice/recommendations, or opportunity of any kind? (If you doubt the basic human need here, please go read Mike's article, now. It may turn you from an anti-social networker to an open social networker.)
G+ is so new. Of course it's going to go through continual changes. Whether it continues to attract audiences, and whether it retains them, depends on a lot of things. (Many more, of course, than I'm covering here.) My favorite anaolgy I've heard so far is "Let’s not judge the cookie by the dough," from Deven Coldewey's post (http://techcrunch.com/2011/07/15/google-one-hell-of-a-trojan-horse/
) where he echoes the sentiments of my first post about G+ (http://bit.ly/o9NC7N
). Will G+ give users more control by letting them sort the feed based on post date vs. comment date? Will G+ give users more control and let users sort by photos, videos, links? Or simply search the stream?
More importantly, will Google use their nearly unmatched strengths (understanding of human language and machine learning) to create features we've never seen before--imagine if G+ could determine the semantic nature of a post, categorize it, and let users follow a subset of topics from a user, instead of an entire feed: (e.g. follow Tom's posts about Google+ and Apple, but not his silly .GIFs).
So to return to where we started -- is "social" in Google's DNA? What does that even mean? I would argue that this means understanding that sometimes humans can do things better than computers. That sometimes when building social software, we need to use social science to understand user desire & behavior. And finally, that sometimes, it's better to think highly of people rather than to assume your product will be too difficult to comprehend. Please give us control Google. If you do, we'll have no reason to complain about your algorithm.