Google Now Reporting Anchor Text Phrases

Hurray! Finally, you can get a report from Google of the top anchor text phrases used when people link to your site. Google Webmaster Central has just announced the new feature. But didn’t Google already report anchor text data? Yes, but only keywords, not phrases. Keywords are mostly useless junk food data. Phrases are datalicious, tasty and helpful. Below, a detailed and illustrated look at what a difference a phrase makes and how to claim your own.

Getting The Anchor Text Data

To access the data, you have to be verified Google Webmaster Central user. That’s explained more here, and it’s easy to do. Note that if you’re verifying for the first time, it may take up to a day for the anchor text and other data to start showing up for you.

Once logged in, select the site you want to view from the “My Sites” screen. When that site loads, choose the “Statistics” tab, then select the “Page analysis” link you’ll see in the left-hand navigation.

Look further down on the page under the “Common Words” section. There are two tables. The one on the right is called “In external links to your site.” That’s the table you want.

That table shows you the top 100 phrases people are using to link to your site. So for Search Engine Land, using the top of that table, here are our top 10 phrases most used when people link to us:

1. search engine land
2. more
3. 25 tips to optimize your blog for readers
4. google releases new link reporting tools
5. wikipedia enters top ten most visited sites
6. will google remain the start page for the
7. george bush is no longer a miserable failure
8. why the seo folks were mad at you
9. google kills bush’s miserable failure search & other
10. click to continue reading

What Is Anchor Text? How People Link To You

Let’s look at the most popular phrase on the list — “search engine land” — to explain more about how all this anchor text phrase stuff works. If you’re a pro, you might want to skip this section. For others, I thought a little anchor text basics might be helpful.

Lots of people link to Search Engine Land from across the web. Anyone can see some of these links to us by doing this search on Google:

link:searchengineland.com

(As a reminder, site owners can see even more links than the search above shows through Google Webmaster Central. My
Google Releases New Link Reporting Tools article from last month explains more about this. But the public link reporting tool is fine for the explanation I’m doing.)

One of the top sites listed linking to us is Graywolf’s SEO Blog, from this page. If I go to the page, I can see him linking to us. The link to Search Engine Land says (that’s what “anchor text” is, the text in a link), “Search Engine Land” like this:

Search Engine Land

It’s not surprising that Search Engine Land is the top phrase used to link to us. It’s our name, so lots of people link to us that way. But now let’s look at something else. “More” is the second most popular phrase used to link to us. Weird, right? Well, chances are there are lots of people linking to our stories and using the words “more” in the link.

How They Link, Not How Many Links, Influences Ranking

As you can see, the data is great information. The text people use to link to you is one of the most important factors — often the most important factor — for how you will rank in Google.

Let me repeat that. The anchor text used to link to your pages often is the most important reason you’ll rank well for particular words.

People still continue to mistakenly think that doing well at Google is about getting as many links as you can. It’s not. It’s about getting quality links from important sites and ideally, very descriptive links — links using the terms you want to rank for in the anchor text.

If you’re trying to show up for some key phrase, the new data will show you if people are linking to you that way, as seen by Google. If not, then you understand that a lot more targeted link building work may need to be done.

Down at the bottom of the table, you’ll see a “Download this table” option. That will let you download the data, if you want to play with it more.

The Old Keyword Table

I started off saying that anchor phrases are better than anchor keywords, which Google started reporting in February 2006. Before today’s change, this is what Google reported about our anchor text as the top ten words:

1. search
2. engine
3. land
4. google
5. the
6. 2007
7. for
8. searchengineland
9. searchcap
10. day

See the first three? On the new report, you can understand that Google was seeing lots of links that said “search engine land” and breaking the link text into the individual words. That was terrible, because it meant you had no idea of the exact order used in your most popular links.

To explain this even more, see the number four term, “google.” Actually, that’s kind of useful to know — that out of all the links out there to our site, Google occurs as the fourth most popular term in them. Potentially, that might help us rank for that particular word. But breaking it apart meant I was unable to see that the word was really part of longer phrases people use to link
to us, like this:

• google releases new link reporting tools
• will google remain the start page for the
• google kills bush’s miserable failure search & other
• google ramps up personalized search
• google‘s agent rank patent application
• google news top of mind even if not
• q&a with marissa mayer google vp search products

If I have to choose, phrases are more useful. But ideally, I’d love both.

Wish List

Now that I’ve gotten one wish — anchor phrases rather than anchor keywords — I want more. My wish list:

• Anchor Keywords: I know, I abused them when starting out this story. But they can be useful to see in addition anchor phrases. So let’s have them both!
• Number Of Links: It’s great that I can see the exact way people are linking to me, but how about the number of those links? Just how many more people link to me using the number one phrase versus the number two phrase? Enquiring minds want to know!
• Who Links? Let me drill down into any particular phrase and see the exact pages that are linking to me in that way. This I want more than anything else, especially because, as I’ll explain, it’s something you can’t get easily from any of the major search engines.

Will we get it? Perhaps, plus maybe some things I didn’t think of.

“There a lot that we can do with linking data that could help webmasters, and we’re looking at how we can provide more information. We aggregate phrases right now, but perhaps we can break those out. We might also show more than 100 phrases in the future,” said Google Webmaster Central product manager Vanessa Fox.

Aggregate Data

By “aggregating” phrases, Fox is talking about how Google will remove punctuation and capitalization, so that all “variations” of a phrase get consolidated into one. Consider people linking to a Star Trek: The Next Generation site in all these ways:

• Star Trek: The Next Generation
• Star Trek – The Next Generation
• star trek, the next generation
• Star Trek, The Next Generation
• Star Trek The Next Generation
• Star Trek | The Next Generation

The same five words are used in the exact same order, so currently, Google considers all the variations to be the same phrase. To do this, it strips out most punctuation (apostrophes and ampersands seem to still stay in) and reports all variations like this:

• star trek the next generation

In the future, potentially those who care could get an breakdown showing exact spellings, punctuation usage and so on.

Internal Links & Subdomains

A few more details about the anchor text data. First, it comes only from external links to your site. Anchor text you use on your own site isn’t counted.

Second, links to any subdomains you have are NOT included in the data. So if you had several sites like this:

• mysite.com
• products.mysite.com
• blog.mysite.com
• videos.mysite.com

Then links to all of them would NOT show up for the mysite.com report. Want to see link data for a particular subdomain? Then you need to make a separate account for that subdomain.

What about this popular situation?

• mysite.com
• www.mysite.com

If these are separate domains (IE, when you enter them into your browser, the domain doesn’t change), then the links are reported separately. The exception is if you redirect. If entering www.mysite.com makes the domain change to mysite.com (similar to what we do here at Search Engine land), then the mysite.com account will cover any www.mysite.com links. FYI, that’s true for any domain or URL you redirect to another domain.

Hunting Down The Links

Remember how I said I couldn’t find an example of the people linking to me and using the word “more” in the links? What a pain! I couldn’t do it. Nor could I figure out why the 30th most popular anchor text link to our site was:

askmeaboutmybra.com

I’ve got nothing to say about my bra. I don’t wear one. Who and why people are linking to us that way is a mystery. Maybe it’s a glitch, but it underscores something I’ve wanted for ages. That’s better search commands allowing me to see exactly what anchor text is used to link to a site.

I’ve especially wanted this as we’ve had various link bombs or Google bombs erupt. Sure, Google’s fixed the Google bomb problem (see Google Kills Bush’s Miserable Failure Search & Other Google Bombs for more on that). But it’s still interesting to research the ways people are linking to sites, the words they are using.

Two major search engines offer anchor text searching. Not Google. Not Yahoo. Instead, it’s Ask and Microsoft that do. However, I don’t think the Ask “inlink” command works properly (Ask is getting back to me on this). Nor is it helpful without a related “link” command. So I’ll focus on Microsoft.

Microsoft added an anchor search command back in June 2005. It looks like this:

inanchor:

If you want to find pages that have particular words embedded in anchor text, then you use that command with the word, like this:

inanchor:miserable

That search should bring up all the pages that have a link containing the word “miserable” in the link. If I wanted to find more than one word in links, I have to do this way:

inanchor:miserable inanchor:failure

That should bring back pages that use both words in links on them — which would be pretty much a way to find all the people who have been playing in the “miserable failure” link bomb game over the past two years.

Unfortunately, the command — like Ask’s — doesn’t seem to work. In both cases, I get back the official President George W. Bush web page. I can assure you, there are no hyperlinks containing the words “miserable” and/or “failure” on that page. In fact, neither of those words even appear in ordinary text.

Let’s pretend it DID work. Remember how I wanted to find the pages linking to Search Engine Land and using the word “more” in the hyperlinks? To do this, I’d need to combine two commands. I’ll bold the two different ones:

inanchor:more linkdomain:searchengineland.com

That would bring back only pages linking to Search Engine Land and which had links using the word “more” in the link text. However, even if inanchor was working properly, this search STILL wouldn’t be enough. That’s because you could have a page that links to Search Engine Land in one place — with ANY text in the hyperlink — and has a hyperlink to some other page that uses the word “more” in it.

I struggle to think of what type of new command could be added to do exactly what I want. That’s what makes the new Google reporting so especially cool. It is the only way you can easily get the most popular anchor text from across the web pointing at your site. If Google adds some of the wish list items, especially the ability to see exactly what pages link using particular phrases,
it only gets better.