While social signals seem to reflect the human value of a page, they could also bring a host of potential biases and distortions.
Despite the recent Panda update, Google still has a problem separating the wheat from the chaff. The truth is that no algorithm can really know what makes one piece of content better than another, quality is human, not mathematical. So Google uses third-party social signals, and its own +1 initiative, to get a sense of what people like.
However, social signals are only democratic. They’re don’t reflect quality, just popularity. And while they may be harder for webmasters to game, they still involve factors that could bias or distort search results. Here are six of them.
Even though I like a link, I might not bother to share it. (Just as I might not bother to review a book I like on Amazon, or a hotel I like on Tripadvisor).
Problem: Social signals only represent the views of those who are motivated to share, who may not be representative.
Often, I quite like something, but not enough to pin my reputation on it.
I once Tweeted a link to ‘the top 50 album covers’ or something similar. In truth, I just thought it was OK, and possibly of interest, so I threw it out there. Someone I respected criticised it, and I realised I agreed with him, which made me feel foolish. Now I think twice before I share.
Or maybe I’m too embarrassed to share. If I found a superlative hair-restoring treatment, would I tell the online world about it, and admit to insecurity about my male pattern baldness? (I discussed the problems for less popular or ‘distress purchase’ brands in my last post).
Problem: Social signals only represent the views of those willing to be publicly associated with a site or page.
Lack of knowledge
Not everyone knows social media well. Many people don’t use it at all. Some might enjoy a piece of content, but not understand how to share it or why they should.
Demographic variations in social-media use must mean that social signals will show a huge bias towards the young, the tech-savvy and the time-rich. So what does that imply for grey-pound brands like Saga or Werther’s Originals, given that over-65s use Twitter and Facebook less than anyone else?
Problem: Social signals only represent the views of those who are competent and confident with social media, not necessarily those best placed to judge content.
Arguably, Google +1 adds nothing. In fact, it may even subtract from the social experience, by adding an unwanted extra option into the mix.
Right now, most webmasters provide ‘Like’ for Facebook, ‘Tweet’ for Twitter and maybe AddThis or ShareThis for everything else. But Google +1 is too big to be sidelined, so we’ll have another sharing option appearing everywhere.
Few people will share everything through every channel; they’ll make a choice. And it’s well known that the more choices people have, the more likely they are to fall into choice paralysis – giving up in the face of a difficult decision. Making your user think too much is never a good idea.
Problem: More opportunities to share could mean fewer shares, even for pages that have real value.
You arrive at a page that you think is OK. Then you see that 10,000 people have liked it on Facebook. Does your opinion change?
The urge to follow the herd, even when the herd is wrong, is incredibly powerful. Psychologists call it the bandwagon effect. Better to roll with the crowd and share what’s already been shared than expose yourself to ridicule by tweeting some unknown link that people might not like.
Problem: Social profile and search profile could form a positive feedback loop. High-ranking pages get shared more, so rank higher, so get shared more, so rank higher…
Google’s idea of including +1 buttons on search results pages is utterly perplexing. What are they thinking? What useful evaluation of a web page can I possibly make from the SERPS? Or am I, perhaps, supposed to return to my search results and vote for a page long after I clicked through to it?
However, all the sharing buttons are prone to hasty usage. It’s just too easy to click them without really thinking about the value of the page you’re on. And of course, you could easily click them without reading at all because you want to look clever or up to date, for example.
A few careless clicks, coupled with the tendency for conformity I’ve already covered, could soon see a mediocre page gathering big props regardless of its value. How many 100-tweet links have you followed that turned out to be mediocre? (And don’t say it happened when you clicked through to this page…).
Problem: Over-encouraging people to share makes it more likely they’ll share prematurely, skewing search results.