[Case Study] SearchMetrics 2013 Most Important Ranking Factors

[oozyluce]

Another victory for the impact of social media sharing as a ranking factor in 2013!

EXCERPT:

FULL WHITE PAPER:​

http://www.searchmetrics.com/media/documents/ranking-faktoren/download-ranking-factor-study-2013.pdf

 

[davidjoy]

very useful! thank you!

 

[Veyron]

Thanks for the great post!

Veyron

 

[SkyrocketSEO]

off-topic- I got a shotgun and rifle and 4-wheel drive, and country boy can survive....
on-topic- Why we see new thread about this every day?

 

[oozyluce]

off-topic- I got a shotgun and rifle and 4-wheel drive, and country boy can survive....
on-topic- Why we see new thread about this every day?

I didn't see any other threads about it but i believe you. i didn't look that far.

This means one thing tho: it's worth sharing!

 

[kelso]

Why we see new thread about this every day?

Because "this" changes every day

 

[SkyrocketSEO]

Because "this" changes every day

No, I ment, this graph is coming back each day.

 

[TZ2011]

Well I see it first time. Thank you oozluce, rep added.

 

[hedgy73]

First time for me too thanks OP. Some quite surprising stats there.

 

[oozyluce]

Well I see it first time. Thank you oozluce, rep added.

You're welcome! It's quite nice to see some familiar faces around here

First time for me too thanks OP. Some quite surprising stats there.

Indeed, it's filled with surprises... In my opinion the on-page SEO data is as surprising as the off-page SEO



...BTW to those who actually read the full whitepaper, anybody understood the "WDF*P*IDF" formula?

It's supposed to bring a new way of looking at the keyword density as well as the page relevance for a certain keyword, and my slow, beer-fed brain is unable to process the information out of it.:albertein

 

[hedgy73]

I agree the on page data is what surprised me the most. Especially the lack of relevance for keyword in url, H1, H2, title and description. They are all at the bottom of the chart, that's amazing.

 

[crazedspyker]

I agree the on page data is what surprised me the most. Especially the lack of relevance for keyword in url, H1, H2, title and description. They are all at the bottom of the chart, that's amazing.

This data is *relative* to their 2012 analysis. (They mention this on their site)

This means social signals *has increased* in importance, but it's still quite negligible. At least this graph shows the trend Google is moving towards.

 

[hedgy73]

Thanks crazedspyker I can see that now for sure, think I need to do some work on my backlink profiles...

 

[oozyluce]

This data is *relative* to their 2012 analysis. (They mention this on their site)

This means social signals *has increased* in importance, but it's still quite negligible. At least this graph shows the trend Google is moving towards.

Read the entire whitepaper dude, these are absolute metrics, not relative to last year. They actually do the comparison with last year's data inside the white paper, but the excerpt chart is not that one!

 

[crazedspyker]

Read the entire whitepaper dude, these are absolute metrics, not relative to last year. They actually do the comparison with last year's data inside the white paper, but the excerpt chart is not that one!

"In addition to providing an overview of the changes in correlation since last year, we have extended our analysis in 2013 to include many relevant new factors."

They are focusing on the changes in correlation, not the absolute values. H1 is still very relevant, but it has decreased slightly.

I don't use any social signals at all for my main sites and they rank just fine, especially post penguin 2.0.

If I followed this graph assuming they were absolute values, I probably wouldn't even be on page 1. Sites don't rank on social signal yet. Maybe it will in the future, but right now it barely makes a difference. A lot of other people have tested this already and can tell you the same.

Then again, even if they were absolute values, they are still just correlations. Correlation does not equal or imply causation. For all we know, sites with social signals might also have another factor in common (with high correlation, not mentioned in this graph, as an example) which is responsible for 95% of the initial correlation (Again this is just an example).

But rep to you because it does show an overview and direction google is moving towards.

 

[learnt]

nice one,thanks for share

 

[Petko]

Then please hit the 'thanks' button, not just trying to increase your post count

Cheers
Petko

nice one,thanks for share

 

[oozyluce]

"In addition to providing an overview of the changes in correlation since last year, we have extended our analysis in 2013 to include many relevant new factors."

They are focusing on the changes in correlation, not the absolute values. H1 is still very relevant, but it has decreased slightly.

I don't use any social signals at all for my main sites and they rank just fine, especially post penguin 2.0.

If I followed this graph assuming they were absolute values, I probably wouldn't even be on page 1. Sites don't rank on social signal yet. Maybe it will in the future, but right now it barely makes a difference. A lot of other people have tested this already and can tell you the same.

Then again, even if they were absolute values, they are still just correlations. Correlation does not equal or imply causation. For all we know, sites with social signals might also have another factor in common (with high correlation, not mentioned in this graph, as an example) which is responsible for 95% of the initial correlation (Again this is just an example).

But rep to you because it does show an overview and direction google is moving towards.

Thank you for the rep But i must rehiterate to you, that those are absolute values and not relative to last year's changes.

In your quote "In addition to providing an overview of the changes in correlation since last year, we have extended our analysis in 2013 to include many relevant new factors." The overview of the changes since last year is in fact present, but it is an addition.


Here's a typical absolute ranking factor graph that they present during the entirety of the whitepaper:

Now, here's the "overview of the changes in correlation since last year" that they're speaking of, for the same data:

As you can see, some bars are greener one one side, meaning that they have observed that this ranking factor has increased in correlation since last year, and some bars are empty with a green border, meaning that they have observed that this particular ranking factor has decreased in correlation since last year.

They are focusing on the changes in correlation, not the absolute values. H1 is still very relevant, but it has decreased slightly.

I don't use any social signals at all for my main sites and they rank just fine, especially post penguin 2.0.

I was surprised at this data too at first glance, but the answer relies in a fact that you have stated yourself: "Correlation does not equal causation"! The reason behind the H1 tag having such a low value is because almost EVERY website from top 1 to top 30 had the keyword in the H1 tag, and therefore there is almost NO correlation with ranking #1 and having the keyword in H1 tag, because everyone is doing it! If every one is doing it from keyword #1 to keyword #30, how can there be a Spearman correlation between the keyword in h1 tag and good rankings? (quite trivial, isn't it!)


Point is, the keyword in h1 is not only very important, it is CRUCIAL:
If you omit to put it, it will have a NEGATIVE impact on your rankings.
If you do put it, the impact will not be positive, but rather neutral, as it is a decisional factor as wether you do get placed into the SERP for this keyword or not.

I hope that made sense

 

[leobar]

thanx for sharing mate. very usefull to know

 

[nomisc]

Thanks.. Its a useful information..