Using an optimised text in the alt attribute of images or in the anchor text in links that are followed improves the search-engine ranking of a page for the keywords used in the alt attribute and anchor text. But what if the link is not followed? Do the alt attribute and/or anchor text in such a link have the same effect or not? Should we distinguish between indexed pages that contain links that are not followed and pages that contain links that are followed, but are not allowed to be crawled by search-engine spiders?
Interestingly the anchor/alt text of nofollow links does count towards the pages SERPs (I was surprised when I tested this). Can't say they count to the same value as a followed link, (maybe it has the same ranking effect as body text, I have no way to test this) but it definitely counts as a ranking factor. Proof try this search in Google- talian "direct line car insurance not good value" This search looks for the word talian (name of a WordPress theme I sell) and the exact phrase "direct line car insurance not good value", (name of an article on another site) so a very specific search with 6 results currently. The first and second results are my sites, the second one uses the text "direct line car insurance not good value" only as the anchor text of a nofollow link within the body of a comment (3rd to last comment). The only way this page could show for this search is if Google counts the anchor text of nofollow links towards the pages SERPs. The result is a little confused since Google has taken talian as a spelling mistake for Italian, the page has an italian translation link to Google languages (top left menu) and this link is an image link with alt text including the word Italian and it's also nofollow. So by luck it shows Google uses alt text from nofollow image links as well (didn't plan to show that with the example ) David Law
Evidence! Good you tested this, although your findings contradict Google's guidelines on the use of the nofollow attribute in preventing comment spam. The nofollow attribute is intended to prevent websites from benefiting from backlinks in spammy comment posts, yet you've shown that such websites do get benefit if the nofollow attribute is applied to links. Is this a flaw in Big G's algorithm?
Think you've misunderstood what I said. The page the link is ON gets the anchor text benefit, not the page it is linking to. So if a comment spammer gets a nofollow link on your blog you could potentially** benefit from the anchor text of their link, but they will gain nothing (other than maybe click thru traffic). ** most likely you won't benefit since spammers tend to post about viagra and blackjack etc... not the content of the page they are randomly commenting on through spamming scripts and so the anchor text is unlikely to be relevant to your pages SERPs, which means your SERPs are more likely to be hurt by this type of linking. I have to admit I've not tested if the anchor text benefits the linked to page since I assumed it wouldn't (that's why it was introduced and find it hard to believe it's a PR (public relations) ploy to shut bloggers up). David Law
So, you're saying that even if the nofollow attribute is applied your website benefits from the anchor text of an outbound link pointing to a website that is topically related to your website?! That's odd. Returning to my original question: there is a link on Big G's page with guidelines to the use of the nofollow attribute that points to an interesting article. This article mentions that there is some controversy about the SEO effect of the nofollow attribute. Unfortunately, no examples are given.
Yes to the first part. Though with comment spammers the anchor text is highly unlikely to be related to your content in which case it will harm SERPs having unrelated anchor text. I was also surprised when I did the tests, expected to find the whole link (including anchor text) would be ignored. The tests I've done only prove it counts, not how much it counts. Could be identical to other anchor text, but then again could be worth less than standard body text. I couldn't find the article your referring to, link? I know there are people who think using nofollow will have a negative impact on your SERPs generally, arguing if your linking to nofollow content (suggesting bad content) that your content will be deemed less important. I find that very hard to believe since nofollow was invented to protect good sites, in my experience (own over 50 wordPress installations, use it as a CMS not just for blogging) that it has a neutral effect on specific pages SERPs. All it does is stop PR transferring from your pages to other pages, so in that respect it's going to have a positive effect since it means more PR flowing through your site. More PR/link juice = better SERPs. David Law
I read your article “Internal links versus Outbound Linking†where you describe the effect of the anchor text of outbound links on the ranking of your own website. I still think it's very odd, because it implies that your website's ranking is improved for non-related keywords or keyword phrases. Obviously, the improvement is close to nil for the popular keywords and keyword phrases used by spammers, but it could rank well for some long-tail keyword phrases (if spammers were using those). It's also odd, because it contradicts the purpose for which nofollow was initially introduced. I did a search and dug up several of the conflicting articles about the effect of using nofollow in an outbound link on the search-engine ranking. In John Andrews' article about what Matt Cutts said at the Domain Roundtable Conference last April, it was mentioned that nofollow "doesn’t flow anchor text", which is what one logically expects rel="nofollow" to do. However, Ben Fisher showed that the anchor text of nofollow links does count for the ranking in Google. Michelle MacPhearson also showed that nofollow links count, not only in Google, but also in Yahoo and MSN. I'm lost as to what the effect is of nofollow on the search-engine ranking of a website. There are just too many contradicting reports. For example about the negative impact of a link to a bad neighbourhood website. In the interview at the Domain Roundtable Conference, Matt Cutts said that a few outbound links to bad neighbourhood websites will not hurt your website's ranking and that the rel="nofollow" attribute is not needed here. But where to draw the line? How many links to bad neighbourhood websites are allowed? I'm certain some webmaster will claim the opposite after having found out that (accidentally) linking to a bad neighbourhood website drastically decreased the search-engine ranking of his website. In an interview on SEOmoz where Matt Cutts talks about nofollow, he said that nofollow can be safely used on internal and external links, but SEO Scientist recently showed otherwise. Does Matt really know what Google's algo does and does not?!
I think you all might be quite surprised if you actually read the principles of 'nofollow' then re-reviewed your posts to patch up the mere coincidences.