A high quality post outlining all details of an ambitious SEO experiment proving that tweets can index a page and make it rank!
With so much controversy about the influence of social media in web results, the only way to figure out whether social media – and in particular Twitter – can influence the SERPs was to actually run a few tests.
If you like this article, you might be interested in some of our older articles on Twitter Tools to Increase Your Productivity, Cool Twitter Tools, Become A Twegend Marketing to Twitter Users, and Twitter Icon Sets.
Great Results? Not Really!
SEO Moz ran a few experiments and they managed to rank well for the term ‘Beginners Guide in SEO‘ arguably because a twitter message got re-tweeted several times. Then about a month later, another experiment was carried out which confirmed the findings of the previous test pretty much as Distilled‘s ‘Excel guide for SEO‘ started ranking arguably again “mostly by retweets“. Following that, SEO Moz did an even more ambitious experiment trying to rank for the super competitive term ‘Jennifer Lopez‘ by asking all their twitter followers to re-tweet a specific message.
Where the Twitter Experiments‘ Valid & Conclusive?
Judging from the comments on the above blog posts, the majority of the SEO community was really fascinated by the results and the high achieved positions. However, the key question here is: would the above tests give the same results having there been no tweets at all?
First of all, one could argue that both blogs (SEO Moz and Distilled) are really powerful and that the specific posts would rank anyway, even without being heavily re-tweeted. SEO Moz is a PR7 site and Distilled a PR6, both having tones of backlinks from thousands of sites, great traffic etc. Also, all posts probably got crawled and indexed instantly as such sites are being continuously spidered.
Testing Twitter Author Authority
Author authority, is definitely a very interesting concept and running a test to measure the impact of it sounded very exciting. However, the test had to be greatly modified in order to reach to some valid and useful conclusions which could be applied in more common occasions as not many own PR6 and PR7 sites with thousands of loyal visitors.
Running the Test in (almost) Full Isolation
Publishing a post on a small, non-relevant and weak site (PR3) where there are NO internal or external links to the page in question, would consist a more interesting test. Also, it would be more interesting running the test sending out a Tweet using a low authority Twitter account like mine (@macmodi). To make it even harder, a was written, making sure that Latin characters would appear only in the page title, and nowhere within the main content. With all the above in place, a really high degree of isolation was expected, unlike to what SEO Moz and Distilled did through their super-powerful and relevant blogs.
I was particularly interested in the following:
1. Would my post get indexed without being linked at all, neither internally nor externally? Obviously, indexation would rely entirely on my low authority Twitter account and my approximately 150 followers, none of whom understands a word in Greek. This way, I would also make sure that my tweet would get re-tweeted a very few times, if any – again unlike what SEO Moz and Distilled did.
2. How long would it take for the post to get indexed?
3. Would my post ever rank for ‘Manolis Vidalis’, which was the only non-Greek text in the entire copy?
Indeed, on the 9th of March 2011 a tweet was sent out, written deliberately in English so there would be a match between the tweet and the page title of the post. It was expected that the tweeted link would get published on Trunk.ly which although consists of nofollow links, still would provide the bots with an.
The first test was unsuccessful as more than two weeks later the post wasn’t cached at all:
There is definitely a reason behind the failure of this attempt:
1. Either Google’s crawler didn’t find its way to my post, neither through my Tweet nor through the nofollow Trunkly link
2. Google found the post but my Twitter author authority was far too low for Google to consider the content of the page worth the be indexed, given that it wasn’t linked from anywhere else
The possibility that the page wasn’t indexed because the content was written in a foreign language should be minimal as Google does index content in all major languages, including Greek. Obviously, running a search in Greek would be pointless given that the page had not been indexed / cached:
Attending Distilled’s SEO Link Building seminar in London (March 2011) was an enlightening experience as great speakers such as Rand Fishkin (SEO Moz), Tom and Will Critchlow (Distilled), Jane Copland and a few other celebrity SEOs gave some truly amazing presentations, with all of them agreeing that social factors already have an impact in rankings. But then, why did the aforementioned test failed? Was my Twitter authority of such low importance that a high quality obituary written in Greek did not deserve a place in Google’s spam-free index? The Web is supposed to be a place where freedom of speech and expression is the corner stone – a place where anyone deserves to publish his ideas crap so others can FIND it.
Interestingly, one week before the London link-building seminar, Matt Cutts had announced at SMX West 2011 that Twitter signals aren’t part of Google’s algorithm? Obviously someone was lying here, was misinformed or was trying to misinform. But would 7-8 top class SEOs lie to the whole SEO community who had paid a fortune to attend the Linklove conference? Probably, the guilty one was to be found at the other side… again! That person was either trying to hide something, or at least disorientate all lazy SEOs who don’t test anything themselves. Martin MacDonald gave the quote of the day saying “An SEO who is not testing, is not an SEO”.
Test No2: Use Twitter To Influence Rankings
Performing another test was mandatory but the next one had to be modified. Still a high degree of isolation was important and the specification of the new attempt had to be modified. It was obvious that the need of a social media expert was necessary as being Mr Nobody on Twitter would not help. The timing was great as Paul Chaloner – a social media guru, was already running his own tests and we joined forces.
Specification of New Test
1. Write a controversial post, which has some search volume
2. Publish it on a low authority and PR site. Again, the site of the samewas chosen because it was:
a. Low authority (compared to Distilled, SEO Moz and Fresh Egg)
b. Low traffic (in case that matters)
3. There are no internal links to the post
4. There are no external links to the post (to the degree that this can be controlled)
5. Make sure the URL won’t appear in the HTML or XML sitemap
6. Publish the post and wait for a few days
7. Paul would send out an optimised Tweet, which would then get re-tweeted by the whole Fresh Egg team.
The post (which wasn’t of the highest quality) went live on the evening of 17/03/2011.
Six days later, the page wasn’t indexed (in case it was we should be looking for another job) and Paul sent out the following tweet, which we was re-tweeted by most of our colleagues:
- Within just a couple of hours the page was indexed!
- Within three hours the page was ranking at #65 for ‘Social media vs SEO’ in google.co.uk
- A couple of days later it reached #25
Questions with Obvious Answers
- Would the page have been indexed without Paul’s tweet?
- Would the page rank that high without Paul’s and his followers’ tweeter authority given the lack of internal juice and the site’s non-relevancy?
If you’d like to find at more about the experiment look at Tweetsepring – a term invented by Paul that clearly depicts the essence of the above experiment.
There can be numerous, but a few a can thing of:
- Optimise and send out a tweet focused on a keyword that needs some shake up. If you manage to get influential Twitter users to re-tweet, then you could certainly see a term like Laser Eye Surgery moving up in the SEPRs. Too good to be true? Just try it!
- Use Tweetsepring to your advantage for sensitive terms where seasonality is a major factor. E.g. valentine gifts, Christmas gifts, mother’s day gifts etc. Or similarly, if I could get Paul, Stephen Fry, or even better, lady Gaga to re-tweet about those amazing wine glasses
- , wouldn’t that help me to hijack the top spot for 1-2 weeks? Staying there for as temporarily as just two weeks in December would make my client super happy!
- One of your pages dropped out of the index? Or maybe got sandboxed or even penalised? Send an authority tweet out and chances are you will get the page back within hours!
I’m sure many will disagree with the above potential uses but before I start receiving libels of negative comments I would advise you to spend the time more creatively and actually run some tests yourselves. As it was earlier mentioned, “an SEO who’s not testing isn’t an SEO”. Speculation alone doesn’t lead to position changes…