The short answer, yes.
Recently I decided to run an experiment to find out how Google was treating social signals – so I set up a brand new domain (http://www.yogamatcompare.com) with 0 backlinks and started acquiring likes and tweets to the domain. Today, just a week after registering the domain, the site ranks for ‘Indian Yoga Mats’ ‘Cheap Yoga Mats’ ‘Yoga Mat Compare’ ‘Luxury Yoga Mats’ and many other keywords, still with not a single backlink (time stamped proof of 0 backlinks here, and here).
Here is a brief timeline and summary of my activity, and number of social signals, with rankings in the table that follows.
| Date + Time | Activity | Indexed? | # of Likes | # of Tweets |
| 9/3/2011 | Bought domain & set up website | No | 0 | 0 |
| 12/3/2011 11am | Acquired first round of ‘likes’ | No | 35 | 0 |
| 12/3/2011 2pm | Acquired 5 more friends into liking the domain | Yes | 40 | 0 |
| 13/3/2011 1pm | Acquired 70 likes | Yes | 90 | 0 |
| 13/3/2011 6pm | Tweeted the site from @marcus_seo | Yes | 110 | 1 |
| 14/3/2011 7pm | Tweeted the site from 3x new Twitter accounts | Yes | 114 | 4 |
| 15/3/2011 2pm | Shared link on 3x yoga-related Facebook Pages | Yes | 117 | 4 |
| 16/3/2011 11pm | Shared link on 1x yoga-related Facebook Page | Yes | 118 | 4 |
Rankings (Google.com)
| Yoga Mat Compare | Yogamatc ompare.com |
Yoga Mat | Cheap Yoga Mats | Luxury Yoga Mats | Indian Yoga Mats | |
| 9/3/2011 | - | - | - | - | - | - |
| 12/3/2011 11am | - | - | - | - | - | - |
| 12/3/2011 2pm | - | #1 | - | - | - | - |
| 13/3/2011 1pm | - | #1 | - | - | - | - |
| 13/3/2011 6pm | #55 | #1 | - | - | - | - |
| 14/3/2011 7pm | #51 | #1 | - | - | - | - |
| 15/32011 8am | #9 | #1 | - | - | - | - |
| 15/3/2011 2pm | #9 | #1 | - | - | - | - |
| 16/3/2011 11pm | #9 | #1 | - | #156 | #62 | #106 |
Analysis
The first thing that is blindingly obvious is that it took under 3 hours for the website to become indexed after receiving the first round of likes – I am currently testing several other domains and monitoring this duration more closely, but I found that rather impressive.
Secondly, you can see a progressive increase in not only rankings but the variety of keywords ranking, as you would likely see in a normal link building campaign, which suggests that if I were to carry on acquiring more and more likes and tweets I could potentially attain competitive rankings for these keywords (although that is not proven, and in my opinion would not be the most effective method of achieving competitive rankings).
So if there was still any doubt over whether Google uses social signals in their search algorithm, mystery solved – they most certainly do. As for Bing and Yahoo, the test domain is not even indexed yet, suggesting that those search engines are not using social signals quite as strongly as Google are just yet.
My Thoughts:
It is my firm personal opinion that within the next two to three years Google will evolve to return results almost entirely based around our personal social graphs. Personalised Search, social signals as a ranking factor, real-time search results and the recent ‘who shared this’ snippets in search results have all been steps towards this. With that in mind, it does not surprise me that Google are using social signals as a ranking factor, as the data that can potentially be captured in a ‘social’ link contains a great deal of contextual information about the person who posted it that a standard ‘link-building link’ does not.
How does this affect SEO? Right this moment I think it’s safe to assume that acquiring likes, tweets and other social signals is a fairly good strategy for not only improving indexation but also rankings.
What do you think?
Another good article Marcus.
One thing not measured that i would find interesting would be the number of pages indexed and the depth of indexation each day.
E.g.
Day 1: 0 Pages indexed, Level Root
Day 2: 10 Pages indexed, Level 1
Day 3: 30 Pages indexed, Level 2
John
Pretty interesting test :).
My question is: is this was a short time effect? I’ve checked now and i can’t see your site. Where did you bought like (you can PM, pls).
Cheers for the comment John!
Agreed, that would be quite interesting to monitor – although I actually only created less than 10 pages so it may be worth testing on a slightly bigger test site. From memory though I believe on 12/3/2011 there were 3 pages ranking (home, hello world post, and author admin), by 14/3/2011 there were 6 pages, and today we have 9 pages ranking.
Need to clean some of them up..
That’s something new to me. thanks for sharing will look forward to plan new projects soon and get better results for them in no time.
@shpyo Cheers buddy :) Were you looking in Google.com or Google UK? The rankings above are for .com (see: http://www.google.com/search?hl=en&source=hp&biw=1229&bih=706&q=yoga+mat+compare&aq=f&aqi=g5g-m2&aql=&oq=)
You can purchase likes from sites like Fiverr.com, although i’m not sure about the quality (I imagine it being a single person with lots of spam accounts rather than someone genuinely asking all of their contents to ‘like’) – with that in mind, I would mix those likes in with genuine likes from your Facebook friends or do some viral blog campaigns so that they appear more natural.
How does one buy Likes?
Good piece of research.
@Marcus: I’ve checked google.pl with lang en, now I see :).
Thanks for sharing :)
Hmm you dont seem to rank for anything other than your domain name, which is unsurpising as there is no traffic for yoga mat compare. Of course putting a link on social media sites will help google index and find your quicker, these websites get crawled every day, google finds the link then links back you. facebook likes wont play a any part in the algo, as they are not applicable to every company brand. How many people are going to like their accountant? None.
@mark Thanks for commenting!
Which search engine did you check rankings in? The test site ranks for the keywords in Google.com not UK – I respectively disagree with your point regarding Facebook not playing a part in the algorithm, as evidently the site was indexed and ranked with only Facebook likes. While you’re right not every company is fit for social it’s my personal opinion that they could be – and those who do will be the ones that will stand out and benefit from this part of the algo. Car insurance is fairly dull sector but they’re all doing great things with social media, as are lawyers and many other traditionally corporate industries and there are a lot of people who find accountancy interesting, therefore they most certainly have an opportunity to do great things.
Looks like it’s been nuked now…..
Yep, just noticed it’s been totally de indexed – I wonder what’s triggered that? It was ranking 30 mins ago, so must have either been read by Google spam team, or had some kind of penalty for suddenly receiving mentions / links from unrelated sites.. weird!
This is really an interesting experiment. Once I checked competition for these, I did not find any competition, can you please share the competitiveness of keywords? As far as my knowledge is concerned, for a non-competitive keyword and which is also your domain name, you can rank as the domain is cached, nothing special about this.
@Marcus Do you think that was just an initial boost for the site ? May be they didnot see any backlinks till now except the social media (likes and tweets)…That would have triggered something…I still believe that there should be a mix of everything!!
Interesting experiment, but I am unable to see the site with Google India.
We have a new site (about one month old) with few incoming backlinks but more than 3000 facebook fans in a really competitive industry and it is doing pretty well in the SERPS, we still don’t know if it is ranking because of the “honeymoon period” or thanks to our facebook page but we seriously hope it is because of our social media campaign, if you’d like more information feel free to write me and I’ll be more than happy to share additional details. BTW we haven’t used twitter yet…
You got nuked cause you bought likes :)
Seriously though, if organic ranking is the goal – the penalty formula remains with bith SEO and Social channels.
Links, Likes, Loogers. If Google hasn’t received your money, all other forms resemble Penalty Material.
It’s possible Google has now sandboxed the domain? Shame if so as it was looking as if this was going to be a great way around sandboxing!
Interesting experiment. I only wonder how sustainable this organic ranking is/was.
I wouldn’t be surprised if it only was a good position for a little while during “real-time search”. The most important issue from my point of view with social linking is quality. The higher the rank of the site linking to you, the better for your own listing, right?
Well, how does e.g. Google defines the quality of a Facebook link or a Tweet? Does he take an influence score (e.g. Klout) into account to check whether the social mention has authority? You know, it’s not because they’re my facebook friend that they contribute to my knowledge. Lots of crap is posted by my friends. “being connected” doesn’t specifically mean that they add value to my journey to get information. (no offence friends!)
Needless to say that buying “likes” works contra-productive. I can easily imagine that Google integrated something into their algorithm to cover this, and that this might just be a reason why the website was punished.
Don’t do tricks, be real, be transparent, be today, be real-time but have patience as well.
I personally don’t believe there is any way whatsoever that Google can tell if likes have been bought or are genuine.
This is an excellent post.It just shows the value of the social networks, particularly facebook.
Like pages linked to a blog, linked to networked blogs ensure you get viral traffic and I see lots and lots of facebook like pages at the front end of google now.
Thanks
Peter
I’ve been pretty sure that Google uses Tweets for a long time now. I think Jen Lopez at SEOmoz is doing a big test to see if she can rank for Jennifer Lopez based entirely on tweets.
I’m not so sure about Facebook Likes. How could the Google bot crawl that? Since Facebook is almost entirely private, Google doesn’t have access to most of the links/likes. Given their competition for market share, I can’t imagine Facebook cooperating with Google.
Saw this on Ed Dales fanpage – was intrigued. Too bad it didn’t last. Maybe the site will come back after google stops dancing around.
@claire – I agree unless an actual human at google stumbled across this blog post and slammed the site by hand.
Marcus, Killer post! I completely agree that Google rankings will move to almost entirely being based off our personal social graphs. Very interesting, thanks again for sharing.
Wow! Scientific experiment. Like the time logs. Gives me heart that social media work is noticed by Google:)Will tweet and campaign for likes with greater fervour now;)
Two main questions remain:
Can google know if likes are authentic? and if it recognizes the likes as bought, won’t it mark the website as a spammer?
Say google can not know if likes are authentic, how long will the efect last?
Interesting articles on ranking with fb likes and tweets…
I’m not sure what happened to your site but I honestly believe this is one of many essential components in obtaining followers and much needed site interaction.
Our latest site on buying local in Wyoming, which is still in the test phase, is beginning to get the “Thumbs Up” approvals we are hoping for. This local interaction is exactly what built steam for one of our previous niche sites.
A simple and yet tastfully designed recipe site with limited original content but with a fb page and a twitter acct, we were able to rank in the top pages of serps. I think the real proof in the pudding is that we only had 100 or so fans that played a part.
When we realized the power of fb and twiiter, we basically shifted focus. Aside from the bi-weekly content we added, social networking and the Thumbs Up were by far the most powerful for us.
We recently sold the above mentioned recipe site for $700. Wasn’t much but we had bigger fish to fry… At the time of sale, we were ranking #2 on page 1 for our main themed kw’s. Several others were also in the top 1 and 2 pages.
I believe this does in fact work, provided you do it naturally and do not force it! Then again, I’m no expert, just a regular Joe sharing my two cents.
Cheers From Wyoming!
Thanks Marcus, I’ll be back to see how this plays out for you. Good luck.
I think there are a number of factors here that aren’t being looked at. First, is Google actually looking at likes? As of last month Mike Cassidy said that Facebook Like’s were not playing into Google’s social search elements, which leads me to question if Google even has the ability to capture this data at the moment.
Another thing to consider is that you are using an exact match domain. Your strongest performing terms are Yoga Mat Compare and Yogamatcompare.com could the domain and on page SEO alone be what is fueling the rankings? I noticed that almost as soon as you were indexed you started seeing some sort of rankings for the term. Have you tried this on a site that doesn’t use an exact match domain?
I think there are a lot of factors here not being looked at that could make this theory hard to prove. I’d definitely like to see a larger data set to see if this holds true across the board.
Is it possible that the Google algorithm used the Likes to create a short term ranking?
Because the ranking was achieved by realtime Likes the subject became more favoured, when the Like links stopped coming the rank status declined also.
It strikes me that a sensible way for such an algorithm to operate in order to keep on top of ‘current’ trends, would be to create short term rank caches. Perhaps, adding authority over time if Like rates are sustained.
If it we true, that would mean that Facebook Like juice harvested in this way would require a regular ‘fix’ of Likes in order to support and maintain good SERPs. Trade off: quick listings, for ongoing work to maintain your Like rate.
But wait, all this Like ranking stuff must be related to the more conventional Google crawler data somehow………??
Google makes my head hurt!
I wonder how contextual the Like’ees were too, it might be interesting to experiment with contextual and non-contextual Like’ees.
Interesting, but reading the earlier post by Tad Chef: http://www.seoptimise.com/blog/2011/03/linking-out-instead-of-link-building-to-rank-in-google.html, on the Yogamatcompare site you’ve linked out from a couple of your blog posts to articles and to your own personal site from the homepage.
Which (if linking out to build rank has credence) really invalidates the test…
Also, the site is built on wordpress? – which automatically pings services when a new post/page is created. Did you block the search engines and then change it to index/follow once you started building likes?
No doubt in my mind social signals play a part in the algorithm however, but to what degree is difficult to interpret…
I’m not surprised by anything that happens at this point with Google. I launched a new site and within 2 days I had the #7 ranking for my one major keyword phrase. Go figure.
Ed: looks like your discovering the same sort of things happening! Thanks for filling us in.
~Cindy
Great experiment.
If facebook likes and tweets can improve one’s ranking in google then businesses with no FB fan page and no twitter account should sign up now. It’s for free anyway.
Social media marketing can help one’s business in a major way.
I love this type of experiment. It’s very interesting to see the impact of Social Media and certainly useful to know that it does have an influence.
This is a really interesting test and pretty much proves the social aspect for Google.
I also agree that in a few years this could be the way things are going.
One question would be how long does the site stay ranked with only likes or tweets, will be really interesting to see.
I have a domain of BusinessPlanAdvice.com and I have been the #1 listing for the term “business plan advice” for years. The problem: only 480 monthly global searches. I’m confident that your primary ranking was for the same reason as my site – the domain name matches your search criteria. Thanks for posting your study – very interesting.
A lot of theories as usual but that makes Internet marketing really interesting. No one knows for sure ;)
Social Media must be very hard for google to weight. Who has authority? Where is the relevance? Right now I believe you get more or less a short boost if you do not keep the likes and tweets on a constant level.
Jonathan, good point on the validity.
Interesting experiment but i think you’d need a month to capture the full affect.
Marcus,
Firstly, kudos on documenting your test and trying to isolate different factors. This kind of blogging (as opposed to just stating ones opinions) is refreshing and needed in our industry.
That said, I think your testing environment is not isolated at all. Firstly, you cannot make any significant conclusions regarding the speed of indexing or ranking of a site that was built on WordPress. As you know, posting a new page or post on wordpress platform site pings a whole bunch of indexing services, Google amongst them, which will significantly speed up your indexing and appearance in SERPs.
Secondly, depending on the authority of your twitter accounts, your tweets were most likely featured on a number of Twitter scraping sites or just sites that show relevant/related tweets. This probably is the reason for quick ranking of all the other pages that were recently featured in the wave of “twitter influences rankings” blog posts, such as the ones on SEOMoz or Distilled. For example, shortly after Jenitta started her test trying to rank for her name “Jennifer Lopez”, performing a search for text featured in her tweet returned thousands of Google results, either from her tweet or from retweets of all the people in the industry. All of those obviously included links.
So to your question whether tweets can get your site ranking, the answer is “probably”. it is not clear whether this is solely due to Tweets or due to the secondary effect of tweets being scraped and featured on scraping sites. it is just that your test is not the proof of this. As for FB likes, i think a better test of showing this would be to take a regular static HTMl site, not an exact-match domain, and just ask people to like it in a controlled manner. I have a feeling that kind of site would not be indexed nearly as quickly as WP site that has tweets supporting it, but I would love to find out i am wrong.
Another point is that GWT cannot be used as an indicator to number of links, at least not in the timeframe you used. I have had sites taken weeks, even months of intensive linkbuilding before those links showed in GWT. A better indicator would be Yahoo backlinks (as you also did) or even Blekko backlinks that can show new links in a matter of hours.
BTW, I see you submitted some press releases too, albeit on March 18th: http://blekko.com/ws/www.yogamatcompare.com+/links+/date
Cheers
Hey guys, cheers for all of the comments – lots of great food for thought :)
I think there are a few things I forget to mention in the post – first of all I did i nfact turn off all of the ping functions in WordPress to ensure that the only way Google could find out about the site was from the social signals. I have ran a test over the past week on another domain with the exact same set up (pings removed) and it is still not indexed with 0 links or social signals.
I agree with the WMT comments – the backlinks in there are not brilliantly accurate, but I just wanted to show an alternative to Yahoo Site Explorer and I have noticed in several instances that WMT tends to exaggerate back link profiles and spot links quite quickly. The main point is that both showed 0.
I’m going to run a few more tests around this and try to isolate them as much as possible – will amend this post with any interesting updates I find!
Cheers,
M
That’s something new to me, I will run some test over some website to find if the social like and shares are better for just backlinks…
Love this experiment! We are trying so hard to improve our search rankings without spending a lot of money.
My take away from the experiment is how quickly you ranked for exact domain name searches. As pointed out in other comments, the results may have been polluted by other factors. Also, while it is interesting learning, your rankings for long tail terms were not likely to drive much traffic. Similarly to a test I ran on blog commenting, no one SEO tactic for manufacturing links is likely to lead to a first page ranking for a competitive term. However, this is great research suggesting the Twitter and Facebook likes are being incorporated into search algorithms.
I think it worked temporarily and your social media links drove up the rank up quickly but didn’t stick. I just ran a report using rankchecker and it was not showing for any of these keywords anymore.
Sigh. I’m not sure if I should be jumping for joy that my social media efforts really matter or throwing water on my blog content efforts. Maybe both.
@Geri Stengel, I don’t completely get what you mean.
It’s an “and story” from my point of view. It’s to cry and to jump for joy.
It’s social media and SEO site efforts (blog is a social medium, right?).
It’s about time. I’ve worked with some old fart SEOs and they refuse to believe in “social signaling”.
It’s time to collect more RT and Likes. Will be useful for future
That’s a really help post, thank you. I never thought they’d put so much weight on social signals. I s’pose your conclusion justifies the creation of great content.
Following on from my previous comment. The answer is yes, you can with ease but with competive keywords / topics you only rank for a short while. Works with Google + too
Wordtracker is very famous in link building. In its link building tool, the only website that pass the Google search engine is Twitter (for nofollow links), not Facebook or Linkedin.
Compared to your analysis, both have huge different. Anyway, regardless of Facebook will influence seo ranking, Facebook does bring traffic and sales. :)
“Google will evolve to return results almost entirely based around our personal social graphs”
Depending on whether google+ starts dominating peoples lives more than facebook will determine whether this prediction is correct or not. Personalised results are only going to show up if your logged into the ‘google cloud’. Even if google+ becomes incredibly popular, there’s still going to be a large market for people who aren’t on board with google+ and therefore tons of reasons to keep doing SEO like you always have.
What a research! Such rare findings are not shared by many. You have a big heart (unlike me) and a strong passion to experiment (very much like me) (Smile). No wonder your blog made me to subscribe to its feeds which proves that I have eyes for quality.
Thank you Marcus.
I totally agree, Marcus. This year, 2012, Google has somehow expressed that social media will now play a bigger role in site rankings. Their own Google+ has been very much improved to attract more people to interact there. Social media is a useful tool as long as you know how to maximize it, along with other proven SEO strategies.
Hi Marcus. Nice article. I know that social bookmarking works but is over abused so I guess Google will discharge this method in the future. There are a lot of software which I’m not going to name here which are endlessly submitting tons of websites from different accounts to social bookmarking websites, so from my point of view social bookmarking websites are not seeing a great future. Just my thoughts.