Monday, November 26, 2007

Google says...

Google says it does, if it sees a link with the nofollow attributed associated with it. Yahoo and MSN are likely to react in a similar fashion, though I haven't yet spoken with them to get exact details since news of their support only just emerged.

If Google sees nofollow as part of a link, it will:

1. NOT follow through to that page.
2. NOT count the link in calculating PageRank link popularity scores.
3. NOT count the anchor text in determining what terms the page being linked to is relevant for.

That's the situation at the moment. Google is going to evaluate how the attribute works, and it could decide to make other changes down the line, it says.

Now let's look at the impact of each action:

1) Not following the link to the page it points at means that potentially, Google might not index the page at all. As said, the more links that point at a particular page, the more likely it is that Google (and generally the other major search engines) will include that page within its index.

The nofollow attribute DOES NOT mean that someone will prevent a page they do not actually control from being indexed, however. If Google finds even one ordinary link pointing at a page, it may then index that page.

In addition, people can submit their pages directly to Google (and most major search engines). So it's crucial to understand that just because someone might place nofollow in a link pointing at your site, this WILL NOT prevent your page from getting indexed.

2) As for PageRank calculations, it's important to remember that PageRank is a pure popularity score (other search engines have similar scoring mechanisms, just not catchy names other than Yahoo's Web Rank). The nofollow attribute means that a link will not be counted as a "vote" in this popularity contest. That can have an impact on ranking, in cases where the impact of other factors beyond pure popularity come into play.

Huh? Say there are two pages, one with a PR score of 6, the other a PR of 7. Even though the PR7 page is more popular from a link counting point of view, it could still get outranked by the PR6 page if other factors such as the words on the page, or the anchor text pointing at the PR6 page, make it more relevant for a particular search.

It's also important to note that nofollow DOES NOT mean you are flagging a link as being bad in some way. Google isn't going to say, "Aha -- nofollow is on this link -- that's a bad link." Or as Matt Cutts, a Google software engineer who helped develop the attribute, said:

"It doesn't mean that it is a bad link, or that you that you hate it, just that this link doesn't belong to me."

Instead, nofollow effectively will cause Google to ignore the link, to pretend it doesn't exist. This also means you shouldn't worry that people will link to you and use nofollow as a way to hurt you -- Google says that won't happen.

3) This leads to anchor text. Generally much more important in ranking well for a particular term is to get the words you want to rank well for put into a link that points at you. With nofollow added to a link, Google won't associate the anchor text in the link with the page the link is pointed at. This, more than anything else, will sour things for link spammers.

Stop Spam? No. A Start, Yes!

The new attribute won't stop link spamming. Many people may still spam simply because they hope human beings will see the links, click through and perhaps convert. As with email spam, maybe only an incredibly tiny number will do so. But since there's no heavy cost to the spamming, that might still be enough.

In particular, much blog spamming is done through automation. So even with the new system in place, some of that automation will keep rolling along. It will no doubt even evolve to spot blogs and other areas that aren't making use of the nofollow attributes, just as smart spammers currently focus on blogs that have been abandoned, rather than irritating active bloggers.

This means other types of systems of blocking spam will likely still have to be used, such as forcing people to input characters from graphics (captchas), registration and so on (The Solution To Blog Spamming at ThreadWatch has a nice rundown on these, and also see Six Apart's Guide to Comment Spam).

While link or comment spamming isn't going away, it's still heartening that it will be less attractive. Site owners have been given an important new tool that lets them control indexing -- something they've not had offered for years. Perfect or not, I'm glad it's emerged.

Vendor Support

Google started developing the idea of a nofollow attribute several weeks ago and quietly shared it with a number of the major blogging vendors. Many of them have now signed on, pledging to support and implement the tag in the future, if they've not already done so now.

As a result, those using systems provided by one of the major vendors such as Blogger or Movable Type (see here for support news) should find that implementing the tags to be associated in links in comments is a matter of flipping a switch. OK, maybe clicking a radio button or drop-down box! Google provides a list of those supporting it here.

Google said it will soon begin talking with other companies, such as those that making forum software. But makers of these packages or any packages could implement support when they are ready.