For years, Google’s
link: command (and see
here) has deliberately failed to show all the links to a website. This came
out of Google’s fear that site owners simply wanted the data to try and
manipulate rankings — which was pretty true.
Instead, they only provided a sampling of backlinks. Today, that changes. Google Webmaster Central is rolling out new support allowing you to view and even download thousands of links to your site (official news here and here).
You still won’t get absolutely all the links Google knows about. In particular, links to any of your pages contained in Google’s supplemental index won’t be shown. But despite this, the amount of links reported will likely be massive compared to the “regular” link lookup command. For example, consider this query:
link:searchengineland.com
That’s showing about 3,000 links to the Search Engine Land web site (note, for some people, you might see no results, due to an apparent glitch). In contrast, the new system within Google Webmaster Central reports to me that I have 57,000 links pointing my way. From 3,000 to 57,000 links — what a different the new system makes!
Why make the shift?
“Webmasters have been wanting more comprehensive link data from us for a long time,” said Vanessa Fox, product manager for Google Webmaster Central. “We created Webmaster Central to communicate better with webmasters and we take their feedback very seriously. We have been looking at ways to provide this data and are ecstatic to make it available.”
Fox added that Google is also more comfortable releasing the data because it is going only to specific site owners, rather than to the world at large. In other words, site owners establish a relationship with Google when they verify through the Google Webmaster Central system. That lets Google show them more information individually but keeps information about all sites hidden from those who might try to misuse the data in some way.
What links can you get and how do you obtain them? The post about the change on the Official Google Webmaster Central Blog is pretty comprehensive (a brief mention on the official Google Blog also here), but I’ll go through what it covers and dive in deeper to the system as well.
The New Links Tab
To access the data, you have to be verified Google Webmaster Central user. That’s explained more here, and it’s easy to do.
Once verified, select the site you want to view from the My Sites screen. Then when that site loads, you’ll see a new “Links” tab when you log in like this:
(Note: if you don’t see the tab yet, keep checking. Everyone should be getting it sometime today).
That tab allows you to see either External or Internal links through January 15 of this month. Google expects to update the data going forward on a monthly basis.
External links are what most people are concerned about — who is linking to your site?
Note that some people operate subdomains. Links from those domains aren’t considered “external.” For example, a site at “mysite.com” might have all these subdomains:
- www.mysite.com
- products.mysite.com
- shoes.mysite.com
- groceries.mysite.com
Because they all use the same root domain, as I’ve bolded, any links from a subdomain to the main domain will not be considered external. Also, any links to a subdomain (such as shoes.mysite.com) will be counted as part of those reported for mysite.com.
What if you want to see links for just a particular subdomain? No problem – you should be able to verify each subdomain and then run a specific report on it.
By the way, Blogspot and WordPress users — if you don’t have your own domain, then any links to your site from another site within those domains won’t be counted as external links. This is true for anyone using the domain owned by someone else. Just another reason to be master of your own domain in both of those places, which is easily done, as our 25 Tips To Optimize Your Blog For Readers & Search Engines from last month covered.
The External Links To Pages Screen
By default, each page of your site is listed by directory, then alphabetically, with a link count next to it. For example:
In the screenshot above, notice how my root URL is listed first, then the pages are listed alphabetically (well, numerically — and if you wonder why we use that number scheme at Search Engine Land, see Movable Type & Rebuild Safe URLs).
Here’s another example of how page listing works:
- mysite.com/
- mysite.com/1page.html
- mysite.com/2page.html
- mysite.com/Apage.html
- mysite.com/Bpage.thml
- mysite.com/subarea/1page.html
- mysite.com/subarea/2page.html
Unfortunately, I think it would be better if pages were sorted by those getting the most links to them, so that you could easily see the most linked to pages within your site. This might come in the future, the Google Webmaster Central team says. In the meantime, there are ways to do this yourself by downloading the data, as I’ll cover in a bit.
To see links for any page, click on the link count number for that page. That will let you drill-down to see just links pointing at that specific page.
Want to warp speed to a particular page without having to scroll to it through the page list? Use the “Find a page” link at the top of the list. That will open up to give you an entry box like this:
From there, just enter the remaining part of your page address. That will shoot you right to data on that particular page.
Listing Links For A Page
However you get to a particular page, drilling down will show you all the links pointing to it, like this:
I’d actually like to see the “first found” date be shown rather than the “last found” date, which I’ll revisit in a moment. But first, an important difference between how the Yahoo Site Explorer tool gives you links versus the new Google tool.
Want to see all the links for an entire site, rather than links to just a specific page? At Yahoo, you set “Show Inlinks” to the “Entire Site” setting to generate a report showing all the actual links to that site (more about this in Yahoo’s help area here). You can then view up to a maximum of 1,000 links online.
Google gives you an overall site link count but does NOT let you see all the links, as you can with Yahoo. You can only see links to particular pages.
For example, look again at this screenshot:
See how the “All Pages” row says that there are about 57,212 links to Search Engine Land? Underneath that is the http://searchengineland.com/ line, with a link count of 25,841 links. Those are the links pointing directly at the home page of the site (http://searchengineland.com/), not ALL links pointing anywhere at the site overall (searchengineland.com).
Downloading Links
What if you do want to see all the links from an entire web site? Easy – download them. Look at the bottom of the table, and you’ll see this:
Note the two options:
- Download this table
- Download all external links
Downloading will let you get far more than the 1,000 links limit that Yahoo lets you view online or download. However, be aware that for some sites, Google will export more than the 65,536 rows of information that spreadsheets like Excel can handle. In those cases, you’ll either have to have another way to massage the data or you’ll need to download data for specific pages, which will reduce the amount.
What’s in the download? Three bits of information:
- Page getting the link
- Link to that page
- Last found date (Yahoo doesn’t show this)
Remember above when I wished for a first found date? That’s because with it, it would be easy to plot a “link growth” chart for a site. For example, consider this:
What I did was download my link data, then use Excel to make a subtotal of the number of links found for each particular day. This starts from about late November when we launched and runs through mid-January. You can see how the links for our new site begin to ramp up and then skyrocket.
The only problem is that these are “last found” figures. That means they don’t properly show growth. Google may find a link on day in early January, then revisit and “find” it again in late January, which would cause the “growth” spurt to happen later than when it first appeared.
First found dates would solve this. It would make it possible to plot the growth of links for a site accurately over a time and provide a sort of link EKG or link trend that would be helpful to site owners.
Indeed, Google could take it a step further. If people wanted to, Google might allow them to share a growth trend with the world. It could even be done by letting people put a badge or widget on their site that when clicked took people out to see their link growth curve, similar to how Alexa stats are provided to sites like this chart for Search Engine Land:
Of course, Alexa’s stats already come under fire as being open to manipulation by site owners. Using link count numbers certainly would also be open to gaming. You’d only need to create a bunch of sites that link to your site to make yourself seem popular.
Last week, Yahoo rolled out its own link badge showing overall link counts, rather than trends, so potentially that gaming will already begin. People might begin building links just to make the Yahoo figures seem even higher.
To really illustrate how the raw link counts tell nothing, consider this tiny snapshot from my own link report:
http://www.techmeme.com/070110/p19http://www.techmeme.com/070110/p20
http://www.techmeme.com/070110/p22
http://www.techmeme.com/070110/p27
http://www.techmeme.com/070110/p29
http://www.techmeme.com/070110/p30
http://www.techmeme.com/070110/p31
http://www.techmeme.com/070110/p32
http://www.techmeme.com/070110/p33
http://www.techmeme.com/070110/p34
http://www.techmeme.com/070110/p35
http://www.techmeme.com/070110/p36
http://www.techmeme.com/070110/p37
http://www.techmeme.com/070110/p38
http://www.techmeme.com/070110/p39
http://www.techmeme.com/070110/p40
http://www.techmeme.com/070110/p41
http://www.techmeme.com/070110/p42
http://www.techmeme.com/070110/p43
http://www.techmeme.com/070110/p44
http://www.techmeme.com/070110/p45
http://www.techmeme.com/070110/p46
http://www.techmeme.com/070110/p47
http://www.techmeme.com/070110/p48
http://www.techmeme.com/070110/p49
http://www.techmeme.com/070110/p50
http://www.techmeme.com/070110/p51
http://www.techmeme.com/070110/p52
http://www.techmeme.com/070110/p53
http://www.techmeme.com/070110/p54
http://www.techmeme.com/070110/p55
http://www.techmeme.com/070110/p57
http://www.techmeme.com/070110/p58
http://www.techmeme.com/070110/p59
http://www.techmeme.com/070110/p60
http://www.techmeme.com/070110/p61
http://www.techmeme.com/070110/p62
http://www.techmeme.com/070110/p63
http://www.techmeme.com/070110/p64
http://www.techmeme.com/070110/p66
http://www.techmeme.com/070110/p68
http://www.techmeme.com/070110/p71
http://www.techmeme.com/070110/p73
http://www.techmeme.com/070110/p74
http://www.techmeme.com/070110/p75
http://www.techmeme.com/070110/p77
http://www.techmeme.com/070110/p78
Those are nearly 50 links to the same article as Search Engine Land from Techmeme. Because of how Techmeme works, even though the article was linked to once in the Techmeme coverage day, the various iterations of the Techmeme home page keep linking to it over and over again.
If you do a trend chart (or even a count), should these “links” be consolidated into one? How about links from a blogroll? And should you count links from sites that are clearly trying to game the system or from untrusted link neighborhoods?
One thought is to perhaps count links (and show trends) only for links above a certain PageRank level, for information shown to the public. That might help reduce some of the gaming but also allow site owners to take part in having fun widgets like Yahoo just rolled out.
Google said they’re considering the link widget idea. It was something I was actually discussing with the Google Webmaster Central team during a visit last week before the Yahoo badge came out, so minds across the web clearly in in the right mindset. We’ll see, going forward.
Overall, I’m thrilled to get this new information from Google. I’ve never felt getting just raw links alone was ever that useful to a site owner. A list of links — especially a list of thousands of links — tells you nothing. What are the most important links, the links from the highest trafficked site, and so on? You need that type of information along with the list for it to be useful.
Google’s new link release doesn’t solve that problem, from a marketing perspective. However, it does set things up for site owners to perhaps get more useful associated data like “first found” dates in the future, along with other ideas that could prove useful in using links as a measure of your site’s health. In addition, it helps solve a long-standing problem for Google, that of site owners nagging that Google doesn’t show them all their links. Well, now you’ve got lots of them, webmasters!
Postscript Barry: Philipp reports that there is a bug with the tool where you can get linkage data for sites that you do not run (i.e. sites that you cannot verify). He proves that he was able to get all of Google’s links but he would not provide the specifics on how the bug works, for obvious reasons. I am sure this will be patched soon, I hope.
Related Topics: Channel: SEO | Google: SEO | Google: Webmaster Central | Link Building: General | Search Features: General













Like This Story? Please Share!
Like Our Site? Follow Us!
Follow @sengineland