|Demystifying Google - The nofollow myth|
| 9:58 pm on Jun 19, 2013 (gmt 0)|
I have long suspected that Google follow "nofollow" links but two bits of news in the last week confirmed it to me at least.
1. Payday Loans - if you track the top SERPS in Google UK several of the redirecting blogs are ranking almost entirely on comment spam with "nofollow" tags in links.
2. PRISM - putting aside the politics what is clear is the NSA rely on Google to tell them what's happening online. As a collector of data to the US Government you need to be crawling everything you can. So you follow all links.
| 8:03 am on Jun 20, 2013 (gmt 0)|
Nofollow is simply about whether a link passes PageRank, not about whether someone else can crawl it.
Google has said, and I believe them, that Googlebot doesn't crawl nofollow links, even for discovery. That said, Google certainly keeps track of those links, and if a page is found by other means (which is more likely to happen than not... nofollow is not a good way to block a page), Google can easily associate the source and destination urls, and it will list the nofollowed links as backlinks in WMT. We've had some discussions on that here, but it's too late at night for me to look for them.
|Payday Loans... ranking almost entirely on comment spam with "nofollow" tags in links. |
Depends on whether the backlink tool you use offers fresh data. Usually there are some hacked .gov or .edu sites in the mix, and the ranking domain or subdomain gets burned pretty quickly. Back in the April 2012 Google Updates thread, on April 24th and 25th, we discussed a specific spam example that was being widely discussed around the web. On page 20 of 30 (set at 30 posts per page), irishsolar and I compare backlinks we see. He was looking at OSE and I was looking at Majestic....
I also discuss the ranking mechanism, which is akin to Google bombing, and usually requires on word of a query phrase to exist on the destination site. It doesn't necessarily have to be body text. Let's not have specifics on the Payday loan sites. My point in mentioning this is that somewhere in the mix, at least in the fresh data, there are backlinks that pass PageRank.
|2. PRISM - putting aside the politics what is clear is the NSA rely on Google to tell them what's happening online. As a collector of data to the US Government you need to be crawling everything you can. So you follow all links. |
Yes, let's keep politics out of this. I think you're making a bunch of unfounded assumptions about what Google does supply to NSA, in my opinion overestimating quite a bit, and then jumping to conclusions from your assumptions. A discussion of Google and NSA is political, and is outside of the scope of the Google SEO forum. Here's a thread which leads to three other WebmasterWorld threads on the topic...
Europe's Commissioner Demands Answers Over PRISM Data Surveillance
| 4:42 pm on Jun 20, 2013 (gmt 0)|
This topic has been discussed many times. Many big sites, especially social sites, automatically put nofollow tags on all external outlinks on the assumption that it discourages link spamming, or at least prevents link spammers from getting any SEO benefits from their activities.
one question that is sometimes discussed is whether Google's algorithm includes nofollow backlinks as part of its evaluation of the naturalness of a sites overall backlink profile. For example, if a site has a thousand dofollow backlinks but only 1 nofollow backlink, it wouldn't look natural and therefore could imply that something is amiss, so that the site isn't as trustworthy as it otherwise might appear to be.
| 4:47 pm on Jun 20, 2013 (gmt 0)|
There's a pretty decent experiment using 100% nofollow links to rank documented here: [socialseo.com...]
| 7:32 pm on Jun 20, 2013 (gmt 0)|
That's very interesting, but there's two things that they could have done (and didn't) that would have made it more convincing:
(a) use anchor text that was completely absent from the site (even the mis-spelling was used in page or url); and
(b) put up a comparison site optimised roughly the same way for the same terms and build no links to it.
Added: there's another interesting point, which is that they did make good comments that added value to popular blogs. So the ranking improvements could be user-metric based rather than link-based i.e. perhaps visitors clicked those links and Google was able to determine that - and perhaps those visitors stayed on their site a while and Google was able to determine that.
| 11:28 pm on Jun 20, 2013 (gmt 0)|
|Google has said, and I believe them, that Googlebot doesn't crawl nofollow links, even for discovery. That said, Google certainly keeps track of those links... |
Not true. I have pages that are both in a nofollow folder and linked to only on my website using nofollow - yet google still has their titles in the index, not just the url. In some cases, when those pages are simple redirects to another website - google has that website's name and description indexed to my url.
| 11:41 pm on Jun 20, 2013 (gmt 0)|
I thought Matt said they still follow no follow links they just don't use the link data in their algorithm.
Google's job is to organize all the information on the web. I am pretty sure they won't be leaving out this big and growing part of the web (ie the nofollowed part)
Blackhatters definitely rank using only nofollowed comment spam but it is in this honeymoon period that it usually works and eventually the sites get thrown out. I am pretty sure although not completely certain that as a long term strategy it is a no goer.
| 12:05 am on Jun 21, 2013 (gmt 0)|
They could ignore any PR juice from the link and not crawl through the link. But they could still use the anchor text to understand what's at the other end. And crawl the page from another url. That kind of ruins the whole original point of nofollow mind you.
I do think they still use the nofollowed link data in some way. We've recently been cleaning up outbound links on our site to 404 pages. Even if Google isn't crawling those links, they still have enough information to know that we link to xxx number of 404 pages, which quite likely doesn't help put us in a good light.