72 replies to this topic

[nethy]

You know, Jill's going to beat me about the head and shoulders (at least) for saying this, but I don't worry about a search engine identifying a site I've worked on. I know I'm improving the site, so I don't feel like anyone needs to pretend I didn't touch it. That doesn't mean I stick my name on the site -- that's just crass -- but I think of it this way: if you walk into a house and the door closes without getting stuck on the frame, the floors are flat, and the place doesn't shake when a dachshund walks by, then that's a sign that someone worked on that house and made it good and solid. And it's a good house. Why should you take that as a signal you shouldn't trust the place?

There is a subtle difference between the existence of an architect in the history of a house and an SEO in the history of a website.

There is no reason to avoid signalling that a house has been designed by an architect in the same way that there is no reason to avoid signalling that a website has been professionally designed or coded or that the copy was written by a professional writer.

Signalling SEO involvement is similar to signalling the involvement of a soliciter in an apology statement, an advertiser in an article or a 'campaign donor' in a politician's policy. Every one of these parties is likely to claim that they are (in SEO terminology) white hat, they follow ethical standard and they do not comprimise the integrity of the article, policy or apology. They are deeply concerned with the effect of these on the recipients (readers, citizens, victims or users in SEO terminology). However, the recipients cannot be blamed for their skeptisism, both because of anecdotal evidence & by an examination of the interests of all parties involved.

[Jill]

Why is Google allowing webmasters to shape Pagerank?

Good question.

[qwerty]

Why is Google allowing webmasters to shape Pagerank?

We've always had some control over it. That's why you don't just slap a link to every page of your site onto every page of your site. Unless it's very small, the site ought to have structure, and that navigational structure affects the PR of internal pages.

I just don't care for this additional way of doing it, because I think it's right that navigational structure should be reflected in internal PR. They represent very similar qualities about a given page. Throwing nofollow onto some links in order to hold a page down, thereby allowing some other page(s) to get more isn't about site structure. It's a way of ignoring it.

Signalling SEO involvement is similar to signalling the involvement of a soliciter in an apology statement, an advertiser in an article or a 'campaign donor' in a politician's policy. Every one of these parties is likely to claim that they are (in SEO terminology) white hat, they follow ethical standard and they do not comprimise the integrity of the article, policy or apology.

Of course I'm speaking about strictly white hat practices. If I were cloaking, I wouldn't be ok with telling a search engine, "I optimized this site. Come check out what I've done." If Gary Hart were an SEO, maybe he'd have done that. What I'm talking about is making changes that are real and measurable improvements. I see no reason to hide that. Maybe an architect isn't the right analogy. Maybe getting some high quality contractor in to redo the house's interior. So now there are no more cracks in the walls, the floors are buffed to a blinding shine, the windows all fit perfectly, and you've got all new, beautiful and energy-saving appliances. The house is worth so much more than it was when you first got it. So when someone stops by and compliments you on what a great place you've got, you can proudly say you hired Company X to make improvements. Why hide it? You should be proud of what you've had done. You made an investment in improving the place, and it's paid off.

[Jill]

I just don't care for this additional way of doing it, because I think it's right that navigational structure should be reflected in internal PR. They represent very similar qualities about a given page. Throwing nofollow onto some links in order to hold a page down, thereby allowing some other page(s) to get more isn't about site structure. It's a way of ignoring it.

Exactly.

[jehochman]

I have it from very reliable sources that rel="nofollow" is just a more granular way of doing the same thing you can do with robots.txt or the meta robots tag. You can make a compliant search engine pretend that the link doesn't exist. Not only does this prevent the contribution of PageRank to the target page, if you nofollow all the links pointing to a page, that page shouldn't get indexed.

Nofollow is a convenience. It won't bite you in the ass. Go ahead and use it. There are situations where you have pages with zero search visible content (e.g. login required, rich media, or AJAX) that are for users only. There's no point in having these pages indexed. Whether you use robots.txt, meta robots noindex, or nofollow on the inbound links is your option.

[nethy]

I just don't care for this additional way of doing it, because I think it's right that navigational structure should be reflected in internal PR. They represent very similar qualities about a given page. Throwing nofollow onto some links in order to hold a page down, thereby allowing some other page(s) to get more isn't about site structure. It's a way of ignoring it.

This comes down, like many SEO things, to a discrepency between what a webaster thinks about a site & what an engine thinks. In many cases a webaster would be right about the relevence of pages. You could say that the privacy policy link at the botom right is about compliance rather then the relevence to users, at best it is important mostly to users who had an interest in other pages/information elsewhere in the site. simply looking at link numbers would not reflect this.

Other times, the 'centrality' of a page cannot bedetermined by a bot. Many sites link home (or somewhere) in the top left logo. This is often kind of hidden and gets few are no clicks cuz its not empasised. Some sites actually have a conversion element up there (an email capture or link or CTA) it is emphasised and it is central, it is where you are trying to send users. The difference is graphic andas far as I know it would be hard for a bot to spot the difference.

But, letting webmasters 'tell' them stuff is not really google-like, is it?.

[nethy]

Why hide it? You should be proud of what you've had done. You made an investment in improving the place, and it's paid off.

Because the involvement of an SEO undermines the statement 'my site is relevent'

When a site is optimised you are trying to make pages appear more relevent. You may (this is the contemporary approach and usually advocated in this forum) follow the practice of making it appear more relevent by actually making it more relevent. I agree that this is the correct approach. However, this is still a means of making a site appear relevent. Making a site more relevent is a way to make it appear more relevent. SEO involvement is going to make the observer doubt the genuine relevence of the page.

I am not saying that sites will be penalised for it. I have no idea. Many here and elsewhere are better equiped to speculate. I am just saying I see the logic in not wanting to be fingered as an SEOed site.

[qwerty]

No, I disagree. I don't perform magic acts on websites to create the illusion of quality, I make them better. Relevance is another matter. A page is relevant to any number of searches, no matter what you chose to optimize it for. But making a site more useful and informative is a good thing.

[Jill]

I have it from very reliable sources that rel="nofollow" is just a more granular way of doing the same thing you can do with robots.txt or the meta robots tag.

Yep, that's exactly how Matt Cutts characterizes it also.

[glengara]

This "granular" sounds quite impressive, is there an alternative term?

[Alan Perkins]

Operates in a more refined/low-level/micro-management kind of way.

BTW, I'd just like to point out that how Matt Cutts is now describing rel=nofollow could be used, we've been describing on this forum for quite some time. Check out this thread, for example.

[qwerty]

That thread was mostly about Google pronouncing that we're all required to nofollow paid links. I only saw one mention (from Dan) in there about using it to manage internal PR. You suggested

Can we agree that, as applied to Google, rel=nofollow is a tool to manipulate Pagerank - nothing more and nothing less?

But that was still in the context of paid links and untrusted UGC, as far as I can tell. But then Dan brings up that

by using nofollow, I can create a PageRank-friendly structure for my site without compromising usability.

[Jimmy Dunne]

I just don't care for this additional way of doing it, because I think it's right that navigational structure should be reflected in internal PR. They represent very similar qualities about a given page. Throwing nofollow onto some links in order to hold a page down, thereby allowing some other page(s) to get more isn't about site structure. It's a way of ignoring it.

Right, that's exactly what it is. Using nofollow in all the right places is the ability to provide 'natural' navigations to humans while still doing what makes business sense for your site. Personally, I don't care much about what I think SHOULD works, I care what actually works. Let's not be too naive here, you could use javascript to do the same thing and somehow that probably wouldn't get the same argument from people(they might just think you are an idiot).

I can understand when people are unclear about what is allowed and no allowed and they don't want to cross the line from white to grey. But it's kind of funny that Matt Cutts is saying it's allowed and within guidelines and there are valid uses for it, and people are saying that they still won't use it because it doesn't seem 'right'. That's not even white-hat, that's like glowing white with a halo I think.

[Alan Perkins]

You suggested

Can we agree that, as applied to Google, rel=nofollow is a tool to manipulate Pagerank - nothing more and nothing less?

But that was still in the context of paid links and untrusted UGC, as far as I can tell.

Er, no, it wasn't supposed to be in any particular context. It was supposed to be in all contexts. That was what I saw as the implication at that time, and now it seems that implication has been expressly admitted.

But then Dan brings up that

by using nofollow, I can create a PageRank-friendly structure for my site without compromising usability

Yep, exactly!

I can understand when people are unclear about what is allowed and no allowed and they don't want to cross the line from white to grey. But it's kind of funny that Matt Cutts is saying it's allowed and within guidelines and there are valid uses for it, and people are saying that they still won't use it because it doesn't seem 'right'. That's not even white-hat, that's like glowing white with a halo I think.

I agree with the spirit of your post. However, a couple of caveats:

• What Matt Cutts says and the reasons why he says it are two different things. A site that uses rel=nofollow on internal links is clearly one that has been touched by some kind of SEO. That's a nice flag for Google, and therefore a good thing for him (in his role) to to encourage.
• Your rel=nofollow policy needs to take more than just Google into account. There are other search engines, and they don't (necessarily) interpret it the same way as Google.

In addition, Matt Cutts suggesting that an attribute be bastardised from the purpose for which it was designed (indicating links you don't trust, or don't know whether to trust) to something else (indicating paid links, or controlling PR flow) is a big deal in some ways. Maybe his hat is getting a bit greyer, rather than others' hats getting whiter...

[DanThies]

A couple points here because they keep popping up:

1) Nofollow is a lousy way to prevent a page from getting indexed, because you can't control every link. If you don't want a page indexed, you use robots.txt or a robots meta tag on that page. You wouldn't attempt to use nofollow to keep pages out of the index unless you literally had no other choice.

2) Anyone who has done SEO professionally has made changes ("improvements") to the structure of web sites to improve SEO results, mainly to get more of the right pages indexed.

3) Nofollow allows you to make some types of structural improvements without taking away links that are useful or legally necessary (privacy, contact, terms, disclaimers), but which are not especially interesting content for search engines to index.

@ Alan, I have to disagree slightly. Nofollow has a meaning - don't follow this link. Like many things on the web, it can be used for more than one purpose. I hesitate to call the use we're discussing now a "bastardised" application, since the original use was to prevent links from being followed and counted. Thus removing one of the incentives for comment spam on blogs.

"Bastardised" was when they decided to insist that we use it to label paid links.

To further abuse the house analogy, PageRank is great at telling you which houses are the most important in the city, but not so good at helping you find the bathroom once you're inside. By putting a nofollow sign on the doors to the closets and basement, we're helping PageRank's wandering drunk find the facilities.