The Value of Links On and Off Your Website

Today, my best practices are being challenged by the link war going on. My best practices still work eventually and over time, but the quick-fix, get cha' to the top on MSN overnight, days are gone.

That's a good thing. That means the search engines are improving their algorithms and preventing webmasters and spammers from going for the quick score. They're actually taking more factors into consideration when ranking a site and it makes sense that a newly launched site shouldn't rank well from the moment it is launched.

How many of us SEO's now own Arelis or some other link development software? WebPosition combined with a few best practices use to be enough, but now it's all about the links; but how many and where should they be on and off your website?

Tools like Areilis are a great way to get your website into trouble. The search engines despise any attempt to manipulate their rankings and tools like Arelis do just that.

And it's been all about the links ever since Google decided they were the best way to rank a website. No one has looked back ever since.

Well, we all know that each link ON your website gets a hypothetical "vote" and the more links you have on a page, the less value each vote has.

Incorrect.

1) Links pointing to your page are a vote. Not the links on a page. They are votes for other pages.

2) This only applies to Google. Yes, links are important to MSN and Yahoo but they don't see them as votes like Google does. That's because what you are describing is how PageRank works. Not how valuable links can be (and why).

We also understand that pointing outside of your domain without a ref=nofollow is a way to "leak PageRank".

To clarify this further, you don't lose PR from the page but you do send less PR to your internal pages. This gives them less for themselves and less to pass on to other pages in your site.

So the general answer to the On-Page question is simple: Fewer outgoing links, with incoming links increased only proportional to the current PageRank.

For example, if the page your editing only has a PageRank of 3, I'd recommend keeping your external links down to 5 or less and your internal links under 10. Use your sitemap page to maintain intuitive navigation for the web crawlers. Now, if your page has a PageRank of 6, you can increase your outgoing links to about 15 and safely have about 50 internal links. You get the picture.

This is not a bad recommendation but too broad in scope. Typically the higher PR pages a important pages in a site like the home page and main category pages. The lower PR pages are deeper subpages within a site.

Typically the main pages are used to channel users to the subpages. These main pages shouldn't have many, if any, external links on them by design. If the goal is to get users to the subpages then external links are contrary to your goals. The subpages at that point should have to content your users are looking for and that may include links to external sites.

Also, you shouldn't be limiting internal links on a page. There's absolutely no reason to do so. Internal links channel PR to your pages which is a good thing. Limiting the number of internal links limits your use of your own pages which is the only thing you have control of. That's just silly. You also lose out on the anchor text of those links which we all know Google loves to use in its ranking algorithm.

Limit outbound links? Sure. Limit internal links? Never.

I've learned that one or two links from PR5+ web page can put you in the number one position for your keywords.

That is only the case if the links are on topic. High PR links from pages that are unrelated to your page will do absolutely nothing for your rankings. It will only increase your PR.

This is also a broad generalization and should be avoided. Do you really think two PR5 links, even if from on-topic pages, are going to get you to rank #1 for "web hosting"? Or "Cancer"? Or even moderately competitive terms? If this was the case everyone would be buying two PR5 links from any page they can. And considering that a PR 5 is relatively easy to get for anyone who tries, I don't see how it can be considered very valuable in Google's rankings (which of course is all this discussion is good for since only Google uses PR).

[*]Make sure the link contains your keywords in the link text

This doesn't help SEO. For internal links it does but for external links it only helps the site you are linking to. The keywords in the anchor text just count towards the keyword density of your page's content. That's it. A drop in the bucket if you will.

[*]The page linking to you has at least a PageRank of 4 or greater

This is the worst advice that gets spit out over the Internet.

1) PR only affects Google so it is shortsighted from the get-go.

2) Every incoming link is a good link. Some are worth more then others but they all add up.

3) PR is much less important then the context and anchor text of a link. I'll take an on-topic PR1 link with good anchor text over a off-topic PR6 link any day.

Limiting which incoming links you seek is limiting your opportunities to improve your sites rankings.

[*]The link is prominently placed amongst text and not lingering in the footer or bullet list

1) Being in a list as no effect on SEO. None.

2) Being in a footer is only bad if the page is long. Then that portion of the page might not be indexed and your link is never found. But being in a footer does not reduce the weight of the link.

[*]If possible acquire links in .ORG or .EDU domains - since they have the highest PR

Very poor.

1) PR is assigned by links, not by domain extension. You don't get high PR just because you are a .gov or .edu. That's very basic knowledge.

2) Let me tell you what the real myth is and then shoot it down. People say get links from .edu and .gov because they are seen as authority sites since the are gov't sites and educational sites. They say the search engines give them a boost simply because of their domains. NOT TRUE.

Web pages on .edu and .gov domains tend to rank well because they contain quality content and many webmasters will link to their content as a result. Both of these are key elements in SEO. Good content and incoming links. Yes, content is king and these sites are further proof of that.

You want to know the real reason why you should get links from .edu and .gov pages? Because it is hard as hell to do. If you get one the odds of your competitors getting one there, too, is small and that link is helping you exclusively (and thus not helping them).

I'd say the first place to start is the Yahoo! Directory and DMOZ. Pay the lousy $300 on Yahoo! - it will pay off.
I sold nearly $1,000 copies of my ebook since I paid for that one little link in Yahoo! Get over your budget and just do it.

Yahoo is rarely worth the $300. You may have had success with it but most people get nothing from it. Most people recommend only doing it if you have a large budget and are in a very competitive environment. Otherwise, it's only worth doing doing if you have a large budget and nothing else to spend it on.

As for Dmoz, good luck getting in. After submitting a site to it you are best to forget about it as it could be week or months to get in...if you ever do. Dmoz's value in SEO also have been reduced thanks to its poor management and tales of corruption.

Hi John,

Thanks for the reply. Its nice to meet others in the industry and discuss our ideas on effective search engine marketing. For example, I most-likely wouldn't have won the 2006 CCA GREAT Award had I not helped so many other entrepreneurs build optimized websites.

I also may not have won Congressional Recognition from Congresswoman Loretta Sanchez had I not inspired the entire falculty at Westwood College to nominate me for the national award I won through Career.org.

So thanks so much for sharing your viewpoint, it just "super" to see that someone is so interested in this absolutely phenomenal marketing niche.

I thought about a few of your comments (thanks by the way), and agree with you on many points.

Tools like Areilis are a great way to get your website into trouble. The search engines despise any attempt to manipulate their rankings and tools like Arelis do just that.

You're right, and that's why I mentioned that I felt uncomfortable about it. I also felt uncomfortable about selling the newspaper at thirteen years old, 5 nights a week while the rich kids played sports. But as a single parent child I had to do the best I could with the tools available to me. For most entrepreneurs this semi-black hat tool is like sending a 13 year old out knocking on doors at 8pm. As I mentioned, pay as you go may be a better choice.

And it's been all about the links ever since Google decided they were the best way to rank a website. No one has looked back ever since.

Your right, it's the unfortunate truth.

Incorrect.

1) Links pointing to your page are a vote. Not the links on a page. They are votes for other pages.

2) This only applies to Google. Yes, links are important to MSN and Yahoo but they don't see them as votes like Google does. That's because what you are describing is how PageRank works. Not how valuable links can be (and why).

I see you had a few late night typos like I usually do. You're correct on the voting language, sorry if my sentence was misinterpreted. I see links like Congressmen or Representatives, each holding a specific number of votes for a candidate, which as you mentioned is for the landing page of the referring link.

To clarify this further, you don't lose PR from the page but you do send less PR to your internal pages. This gives them less for themselves and less to pass on to other pages in your site.

Thanks for clarifying.

Typically the main pages are used to channel users to the subpages. These main pages shouldn't have many, if any, external links on them by design.

Agreed.

If the goal is to get users to the subpages then external links are contrary to your goals. The subpages at that point should have to content your users are looking for and that may include links to external sites.

For those reading, John made a GREAT point here: Limit outgoing links on your homepage, but I still recommend only pointing to sites with PageRank. Outbound links pointing to a PR0 could hurt you, especially if the page your pointing to has been banned.

Also, you shouldn't be limiting internal links on a page. There's absolutely no reason to do so. Internal links channel PR to your pages which is a good thing. Limiting the number of internal links limits your use of your own pages which is the only thing you have control of. That's just silly. You also lose out on the anchor text of those links which we all know Google loves to use in its ranking algorithm.

Limit outbound links? Sure. Limit internal links? Never.

We'll have to agree to disagree agreeably on this one. I took one of my client's site from virtually nowhere to top 5 results by simply limiting the number of internal links and beefing up the content on the few pages they had left. What I should have done, which would have been even more effective, would be to include a sitemap, because as you mentioned PR decreases with depth. A sitemap keeps everypage linked in a way to have less depth and higher respective PageRank.

That is only the case if the links are on topic. High PR links from pages that are unrelated to your page will do absolutely nothing for your rankings. It will only increase your PR.

You're right, which is why I included my list in the post.

Do you really think two PR5 links, even if from on-topic pages, are going to get you to rank #1 for "web hosting"? Or "Cancer"? Or even moderately competitive terms?

Absolutely. A friend of mine, Lance Larsen started OC Hosting right out of college. The college was so proud of him that they wrote an article on OC Hosting and linked to his websites (the college had a PR8 on the article page). I ran a backlink analysis on his site not too long after they opened and found about 6 incoming links (5 of which were NOT on topic and were unrelated). The one on topic with the Link Development best practices observed was from the college.

This doesn't help SEO. For internal links it does but for external links it only helps the site you are linking to. The keywords in the anchor text just count towards the keyword density of your page's content. That's it. A drop in the bucket if you will.

Right. That's why my comment was under the subheading "External Links".

This is the worst advice that gets spit out over the Internet.

1) PR only affects Google so it is shortsighted from the get-go.

2) Every incoming link is a good link. Some are worth more then others but they all add up.

3) PR is much less important then the context and anchor text of a link. I'll take an on-topic PR1 link with good anchor text over a off-topic PR6 link any day.

Limiting which incoming links you seek is limiting your opportunities to improve your sites rankings.

No need to be rude John.

Link aging is an important factor in consistent placement in the search results, so I agree with what you're saying as it applies to fresh linkage. However, I stand by my PageRank comment. Case in point, my pal Lance's nearly instant ranking improvement with one link from a .edu.

Consider this: You have 2,000 incoming links. 1,950 of them have your url in the link text (www.somewebsite.com). 50 of them have your keyword in the link text (that's 2.5%). You competitor has 50 incoming links with 40 of them containing the link text (80%). Who do you think will rank higher (if all else is equal)? My experience has been the 80% percent company.

Being in a list as no effect on SEO. None.

2) Being in a footer is only bad if the page is long. Then that portion of the page might not be indexed and your link is never found. But being in a footer does not reduce the weight of the link.

Good point, but my experience has shown different results. I started two similar websites with completely different topics. I placed 3 links to each externally, one in a list, the other emerged in paragraphs. The winner, the one wrapped within a paragraph. But maybe you had a different experience?

Very poor.

1) PR is assigned by links, not by domain extension. You don't get high PR just because you are a .gov or .edu. That's very basic knowledge.

Call it what you will, but links I've added from .edu's and .gov's have paid nice dividends.

Yahoo is rarely worth the $300. You may have had success with it but most people get nothing from it. Most people recommend only doing it if you have a large budget and are in a very competitive environment. Otherwise, it's only worth doing doing if you have a large budget and nothing else to spend it on.

The problem those people you're referring to might just be that they failed to include their #1 keyword phrase as the title of their listing. Two recent sites I built were #1 on Yahoo! within a week and already converting.

Thanks again for the constructive criticism. Not exactly how I'd approach a subject if I disagreed with a viewpoint, but I do value your input.

Thanks Again,

Glad you took the time to make your original post and to construct a reply to mine. It's nice to have a good conversation about SEO. :)

We'll have to agree to disagree agreeably on this one. I took one of my client's site from virtually nowhere to top 5 results by simply limiting the number of internal links and beefing up the content on the few pages they had left. What I should have done, which would have been even more effective, would be to include a sitemap, because as you mentioned PR decreases with depth. A sitemap keeps everypage linked in a way to have less depth and higher respective PageRank.

I'm going to have to say this is just a case of jumping to conclusions. A lot of factors affect rankings and they are always in flux. I'm willing to bet other factors caused this. The only way to say for sure is to attempt to recreate. I doubt either of us is interested in that.

No need to be rude John.

That was not intended at you specifically. It's just a common piece of misinformation that is out there and it unfortunately confuses a lot of people.

I stand by my PageRank comment. Case in point, my pal Lance's nearly instant ranking improvement with one link from a .edu.

Like above, there are so many other factors affecting the SERPs to say one link from a .edu is that powerful is quite a stretch.

Consider this: You have 2,000 incoming links. 1,950 of them have your url in the link text (www.somewebsite.com). 50 of them have your keyword in the link text (that's 2.5%). You competitor has 50 incoming links with 40 of them containing the link text (80%). Who do you think will rank higher (if all else is equal)? My experience has been the 80% percent company.

That is correct and is a fundamental principle in seeking links. The anchor text is one of the two most important factors (the other is that the page which links to you related to your page. If so, it's a very good thing, if not, the link loses most of its value).

Call it what you will, but links I've added from .edu's and .gov's have paid nice dividends.

The links helped, but the fact that they were from a .edu or .gov did not.

BTW, One of Google engineers just confirmed that .edu and .gov sites don't get any special treatment and their links are no different then any other. :)