Use Google Verbatim to Assess Anchor Text Impact.

Submitted by Andy Langton on Wed, 30/11/2011 - 10:02

This is one for the SEO techies and link builders. Short of time? The Google Verbatim option appears to strip the relevance usually provided by (some?) link anchor text. You may find this useful in analysing rankings.

I came across an interesting quirk of Google's new "verbatim" search option today. In addition to removing various query rewriting and personalisation options, it also appears to remove the relevance usually passed via link anchor text. This can give some insights into rankings which SEOs may be interested in.

I'll assume that you know about anchor text and its influence on relevance. The one line recap is that the text used in a link increases the rankings of the destination page for those words. It's why there's an entire link building industry :)

A search that highlights this factor is "click here" - since the sites that rank are not those that have lots of text related to “clicking there”, but because there are many links to them using those words. Here's the current top 5 in Google UK when searching for “click here”:

Adobe Acrobat is the first result – because of all the links prompting users to “click here” to install their PDF reader (aside: don’t use generic text in your own links!). The text on the page doesn’t reflect this.

If we search for “click here” with verbatim enabled, we get rather different results:

Top 5 Google UK results for [click here] with verbatim enabled

Now, the focus shifts to sites with greater textual relevance. The implication being that some of that anchor text relevance has been removed. We can use other specific examples to show that individual links are discounted with verbatim enabled, for example one of my old anchor text tests:

http://www.google.co.uk/search?q=%22jolly+good+internet+marketing%22&pws=0

vs.

http://www.google.co.uk/search?q=%22jolly+good+internet+marketing%22&tbs=li%3A1&pws=0

In the second example, the anchor text is no longer passed to the Receptional site. This suggests that certain links are being removed. Perhaps it’s the links where the anchor text is not present on the landing page, which would make sense. Partially matching anchors might still be passing value.

So, what use is this? Knowledge is power, of course, and this provides a way of assessing the impact of anchor text. Put another way, you can see if you’re being outranked by links, or whether it’s your links that are providing your rankings. Fill in the blanks for yourself! Comments or other insights are welcome, of course – so far my testing of this has been fairly limited.

About the author: Andy Langton

Andy Langton is Receptional's Head of Organic Search and Chief Technical Officer. Follow Andy on Twitter @andylangtonuk and engage with him on both Linkedin and Google+. Andy also has his own website offering both technical and SEO support and advice.

Position: Head of Organic Search / Chief Technical Officer

Andy Langton's blog

Comments

Incidentally

Submitted by Andy Langton on Wed, 30/11/2011 - 10:07

Incidentally, terrible snippets for Adobe's results, eh?

Interesting post, thanks! I

Submitted by Rachel McCombie (not verified) on Wed, 30/11/2011 - 11:27

Interesting post, thanks! I wrote about Verbatim recently - http://www.seoptimise.com/blog/2011/11/introducing-the-new-verbatim-sear... - but it didn't occur to me that it might provide insights in this way. I did run a lot of search queries to see if there was much difference between with and without Verbatim and most of the time saw very little if any difference. I was a little confused about your conclusion - so what would it mean if the results were the same with or without Verbatim?

Hi Rachel, a few explanations

Submitted by Andy Langton on Wed, 30/11/2011 - 11:37

Hi Rachel, a few explanations which may help!

I used "click here" as an initial example since no-one optimises for this keyword, and it often isn't used on the pages at all. So, it's a good "control" to illustrate the impact of link value/anchor text.

The final "jolly good" test shows that the difference in results was as a result of the removal of value from an external link - as there were no other relevance factors to consider.

Verbatim's intent is to stop query rewriting, so What You Search For Is What You Get (WYSFIWYG - did I just invent that? ;). However, if Google is also removing link value then the search could show us where a site is ranking because of links - or otherwise. There's likely to be less rewriting for competitive queries, since everyone will have on-paged their sites to fairly high levels.

For a real world example, let's use "SEO"

www.google.co.uk/search?q=seo&pws=0
vs
www.google.co.uk/search?q=seo&tbs=li:1&pws=0

The biggest difference I see is Google's SEO Help article pop in with verbatim enabled. So, the initial suggestion is that Google has the onpage right, but they don't have enough anchor text links compared to their competition. So that could be where there should focus effort to break into the top 10.

Hope that clarifies a little!

Ah, I see - thanks! I will

Submitted by Rachel McCombie (not verified) on Wed, 30/11/2011 - 17:24

Ah, I see - thanks! I will try it out on some of my clients' keywords and see if I spot anything interesting :)

Exactly - give it a go and

Submitted by Andy Langton on Wed, 30/11/2011 - 18:07

Exactly - give it a go and see what happens! I think every SEO should have "playtime" with Google results :-)

Do you know how much work you've just created?

Submitted by Graham Loosley (not verified) on Thu, 01/12/2011 - 08:47

Andy,

Good find, thank you. There I was bemoaning the loss of the + operator and the - operator going wonky and verbatim seemed a poor substitute.

So now you've given us lots and lots of rankings to test.

I'd love to see the inverse of this query i.e. the rankings with the on-page factors stripped away and just the off-page driving the algorithm.

I suppose we can work it out for ourselves by taking the standard rankings, comparing to the verbatim rankings and the difference is the on page?

Graham

A note of caution

Submitted by Andy Langton on Thu, 01/12/2011 - 11:01

Two factors you need to be aware of if using this data:

- Verbatim does have other effects! So, you are unlikely to see the exact result set a general-purpose searcher would see. This will be worse for obviously expanded queries - plural vs. singular, for instance

- This method doesn't seem to remove all anchor text benefit. We would need further testing to assess exactly which links are discounted

@Graham - don't forget that there's is also data to be had using Google's advanced operators. See in particular intext: and inanchor:

However, I've found results for those operators to be fairly spotty. The data is still well worth a look though. In addition, we may find interesting results by combining verbatim with the older operators. So, we may be able to approximate an "on page" search with the below:

http://www.google.co.uk/search?q=intitle%3A"seo"+intext%3A"seo"&tbs=li%3A1

And an off-page evaluation:

http://www.google.co.uk/search?q=inanchor:%22seo%22&tbs=li:1

One of the issues is that there is an amount of "bleed" across the indexed data, so the results are far from precise. That said, if you want an insight into how Google's indexing (not spidering!) process works, advanced operators are a goldmine. But that's a complex subject; perhaps one for another post?