Why are the links reported in Google WMT different from reality?

[RichTC]

I have noticed recently that the reporting fuction in WMT that tells you who is linking to your site is no longer providing the complete picture.

As an example one site i work on is showing in WMT as having 64 links to it from other sites and details those websites. i have printed it off.

Meanwhile on both Yahoo and MSN they report the links found of between 310-325 and have printed that off.

On closer inspection many of the sites that are linking in that are NOT showing on WMT are good sites that would without doubt have been crawled and indexed by google. So why wouldnt they be showing? They dont have no follow or no index attributes and a good time period has gone hence Google would have found them imo.

Another sites i work on has a similar experience. It has 47 links in WMT yet 172 Yahoo/ MSN.

If they are not in WMT im assuming the site is not gaining any credit from them in Google.

Anyone else experienced this?

Rich

[TheMadScientist]

They only show a sample of links using either the link: operator or in WMT ... Been that way in WMT as long as I can remember ... Unless they've changed something I haven't heard about, that's still the way it is.

[robzilla]

If they are not in WMT im assuming the site is not gaining any credit from them in Google.

You can assume that all links crawled by Googlebot are weighed individually, regardless of whether or not they appear in the backlinks report. If Google were to show you all the links that are not somehow discounted by the algorithm, and filter out all others, that would obviously offer a bit too much insight to webmasters (link spammers in particular). As would providing webmasters with a full picture of their backlink profile. So, indeed, it's only a sample.

That also works the other way around, of course. As Matt Cutts wrote in 2007: "Do not assume just because you see a backlink that it's carrying weight." [mattcutts.com ] On the amount of backlinks reported, he then wrote: "The backlink tool doesn’t show 100% of the backlinks from Google yet, but I expect the number of links that are available to grow." They are now showing more backlinks than they did back in 2007, but I doubt they'll ever give you a complete report.

[tedster]

Some sites I work with go the other way. WMT shows a huge number of links that are far above even what Majestic or Linkscape show - and that NEVER used to happen. What's more those links never seem to actually be on the web when I try to check them out - in fact the URLs are often 404.

I'm wondering if the whole thing isn't strongly skewed by a spammer scheme of some kind - publishing link pages only until googlebot grabs them, then freeing up those resources.

[g1smd]

WMT seemed to provide reasonable results until February 2010.

Since then the reports seem to be mostly garbage, very very wrong.

[lucy24]

I'm wondering if the whole thing isn't strongly skewed by a spammer scheme of some kind - publishing link pages only until googlebot grabs them, then freeing up those resources.

But who benefits? People don't get anything out of outbound links do they? I don't quite see someone sending out e-mails saying "Last month I gave you eighty-five bogus links. Pay up or I'm telling Google they were fake."

I don't think google can tell the difference between real links and spurious ones. If I post a comment to a blog and dutifully fill in the box asking for my www site, that's not a real link. If someone's Facebook page says "Looka this cute e-book I found", that's not a real link. If someone posts to an indexed forum with a question about a specific page, that's not a real link. But there they are on the list. I guess I should be glad they're not counting hotlinked images. (Those auto-generated analog stats do. Grr.)

This kind of thing is easier to track if your total numbers are low enough that you can eyeball the list of site names and say "Nuh-uh, never heard of 'em" or "Oh, come on, that's not a link, I filled out that form myself."

[tedster]

People don't get anything out of outbound links do they?

Even in the earliest algos, getting classified as a hub (a useful reference source of outbound links) held value. And Matt Cutts has made several comments that indicate Google appreciates sites that link out.

A few years ago I did a bit of an experiment when I helped a friend build a website for her one person business. I took on the discipline of including at least two outbound links on each article page. The result was that she ranked well within weeks, never hit the sandbox (at that time new websites had one heck of a struggle for a few months), and today has succeeded beyond what she hoped for.

Even today, she has backlinks from 92 domains, far lower than her competition, but she often outranks them. I don't pretend to have a formula for this factor, but based on that one experiment, I'd say outbound links have some helpful effect.

And when I described the possibility of a spammer scheme, I also have in mind sometimes (maybe even a lot of the time) we stumble over spammer experiments rather than something that is actually working. People can only find loopholes in the algorithm if they experiment heartily.