40 is the new 30, green is the new black, pale is the new tan, hipster is the new homeless, & helpful is the new viral – but when it comes to search engines and ranking algorithms Likes are the new Links.
Unless you’ve been too busy actually doing work to read the latest SEO blogs, you know that Google recently announced their plus one initiative while Bing was busy inking a deal with facebook.
While these are two huge undertakings in their own right, I think they’re just the start of a new trend in search engine algorithms – a trend that you most assuredly want to be in on from the beginning.
The paradigm of the web is shifting. Search engines are starting to move away from links and more toward other signals; and you know what? It makes perfect sense.
Topeka Meet the new boss, same as the old boss (if you don’t care about the history of Pagerank and how it works, you can skip this part)
Think back to the late 90’s when Google first came about. They turned search sideways by revolutionizing the paradigm. Instead of looking at meta tags and page content alone, Google used links. The original Pagerank algorithm was based on the model of academic citation, or what some of us computer scientists call the wandering drunk algorithm*
When Page and Brin noticed that the most important academic papers often get cited the most by other research papers they quickly realized that the same model could apply to links on the web. The most linked to sites (like the most cited papers) must be the most important. Suddenly, Google’s search results became much more useful and the web was a better place.
So if links are so awesome, why change?
Because the web is evolving. It may not seem like it, but when Google launched there was no Facebook, no Twitter, no Myspace, no Linkedin, no Foursquare, hell there wasn’t even a Friendster. Links were the signal back then. Now, we’ve got a whole lot more signals.
Not only that, but links have become very easy to manipulate. You can buy links pretty easily, and it’s trivial for anybody with a week’s worth of PHP knowledge to write a script to start automatically building them. (trust me on this one!) Quite simply, It’s very easy to manipulate links with hardly any repercussions.
So how are likes and pluses different? Can’t you buy them too?
I guess you could, theoretically, but likes and pluses are different from links in one key area: they involve your friends. See, if somebody pays me to put a link on my website it doesn’t really affect anybody. Most of my visitors will never see it. That’s not the case with likes or pluses.
If you pay me to like or +1 something, my friends see that. If it’s a page about buying pills online while refinancing my house and playing poker you can bet my friends will not only see it, but they’ll probably be very annoyed at me for posting it. How many links like that do you think I could share on Facebook or twitter before people started unfriending and unfollowing me? My guess: not many.
In fact, I tried it with sponsored tweets a few months ago. It only took 2 sponsored tweets per day to start losing a significant number of followers. (strangely, when I had it at 2 per week or less, nobody noticed)
In math terms: the Limit of Friends as paid posts goes to infinity is 0. In layman’s terms: the more sponsored posts I make the less friends I’ll retain, thus the less valuable my sponsored posts become. That’s a key difference between links and likes: The opportunity cost of spamming my friends is too high. The more I do it, the less valuable it becomes
And that’s why likes and plus ones are going to be more valuable than links. Sure, links will never die. They’ll probably always be a part of ranking algorithms, but likes and plus ones have so much more potential for making search more relevant.
The next level of search
It’s no secret that Google and Bing have been taking search in a personal, customized direction. We’ve seen results catered to location, click history, preferences, and even to my twitter feed. +1 and likes are just the next level.
Likes and pluses will never completely replace on-page relevancy, but they can certainly add to it and help refine my results in a way that links can’t. If I’m searching for something, what’s more relevant to me? The page with the most links, or the page that 20 of my friends have shared?
And that’s just a start. Imagine the possibilities of search results if we add in other signals like facebook and foursquare check ins. Say I’m in Seattle for business and am looking for a good bar. I’m in the mood for a Guinness so I pull up Google local and search “Irish pub.” What’s more useful? The closest bar to me, the one several of my friends “like” on Facebook, or the one where several of my friends have actually checked into (or gasp, left a review for!) in the past month? I’m going to want the bar where my friends actually went when they were in town (assuming they liked it)
That’s the future of search: It’s going to be a lot like dating. Pretty soon, having on-page content and links won’t be good enough. Just like having good looks or lots of tattoos, content and links will get you an invite to the party, but that’s all they’ll get you. What happens once you’re there all depends on how likable you are and what your friends have to say about you – and that might leave several sites (and SEOs) standing alone by themselves at 1:50 wondering “where do I go now?”
* What’s the wandering drunk algorithm? Let’s assume Ryan has too much to drink at SMX advanced and although he finds his way to his hotel, he can’t remember what room he’s in – so he starts randomly trying rooms with his key.
If we count how often he tries each room, (again, assuming randomness) the rooms he tries the most often would be the most “relevant” rooms. That’s kind of how pagerank works.
(trust me on this, I’ve tried this theory several times in various parking garages)
Comment by Ryan — April 5, 2011 @ 12:14 pm
I agree that sentiment analysis is going to go from being a discrete factor into a much bigger factor in the algorithm for SERPs. I don’t think enough people realize the sheer weight of implications of this feature. I think Google Hotpot will become a top contender in the next few years. Both Hotpot and this +1 is Google’s way of competing with Facebook and the likes of Yelp (no pun intended).
The switch from links to likes will be gradual, but it’s definitely happening!
Comment by Pashmina — April 7, 2011 @ 9:34 am
Some great thoughts there Ryan. I personally have my doubts over how well +1 will take off with users.
Do you think maybe Google have got it the wrong way round? With a LIke button you can read the content and evaluate it for yourself from which you can then Like it. Until +1 for websites comes along you only have the opportunity to +1 content before you see it. So there’s a couple of points that raises
1 – are you going to +1 a search result before you read the content?
2 – will a user even both to perform that extra click and +1 a search result before clicking on a listing?
3 -Facebook has a platform to which a Like is registered and seen by friends. Google seem to be doing the same thing but in reverse. +1’ing something before there is any platform for it to register on.
Just a few of my initial rumblings. Ravings of a mad man as usual! 😉
Comment by Rob Kingdom — April 14, 2011 @ 12:58 pm
You’re 100% right Rob. I share the same concerns you do. I think the “+1” button for your own website is crucial to this. Until then, it’s going to be SEOs and mechanical turk people +1ing their own stuff or stuff they were paid to +1
There’s no way my mom or dad is going to visit a website, complete their business, then hit the back button or re-search for it to +1 the page.
Comment by Ryan — April 18, 2011 @ 3:29 pm
yer this thing to me looks like it’s going to be too easy to game! how many people have already started selling +1’s? why not set up hundreds of Google Profiles and outsource +1ing loads of your content. there must be a fairly trouble free way of automating that process as well.
be interesting to get your thoughts on what impact duplicate content could have here Ryan. now i am not really a big believer that duplicate content is a huge problem but here’s what i’m thinking –
if you have a site managed by a CMS that through no fault of it’s own happens to create duplicate pages (I am of course talking about a client site! ;)). Google might start to have problems figuring out which page to index for particular keyphrases. What if Google listed a duplicate page for a keyword or phrase you are optimising for and that content starts getting +1’s? does that make the authentic page less valuable and less likely to rank? is come smart ass going to figure out how to 301 +1’s?
the more and more thought i give this then the more potential problems i seem to find!
Comment by Rob Kingdom — April 19, 2011 @ 4:03 am
I used to drop by this site all the time, about a decade ago. Glad to see your still around. Although the content seems to have changed I look forward to reading your stuff. It was always good.
Comment by SpurtReynolds — June 7, 2011 @ 11:50 am