Category Archives: On-Site Authority Optimization

Reuse URLs to Preserve Authority

As a final note on the topic of on-site authority, whenever possible you want to re-use existing URLs that already have links to them, rather than building new versions.

For example, if you have an article on the best vacation spots of 2013 that got some links and attention — reuse that page when you do the 2014 article! Rather than making a whole new page and starting over from scratch, just replace the contents with the 2014 content. Then your new article is already starting out with a bunch of links pointing to it and a history of good user metrics, making it more likely to rank.

If you really want to preserve the old content, make a new article for the old stuff and backdate it!

Obviously this technique only works if the old content is no longer attracting visits and links. This is commonly used in the ecommerce industry for Black Friday sales: sites like and don’t make a brand new URL every year — they use the exact same Black Friday URL year after year so that it can continue to accumulate more and more and more links (which makes it increasingly difficult for anyone just entering the game to compete for rankings).

You should think ahead when using this technique and make sure that your URL does not include date info. Amazon surely doesn’t want its Black Friday page to live at, for example.

Nofollow Links & When to Use Them

We’ve mentioned nofollow links before in the authority overview, and it’s about time to explain them in more detail. Nofollow is just an attribute you can give to any link you create to prevent that link from passing any authority or ranking factors. The html for a normal link looks like this:

<a href=””>Click Here</a>

That is the standard, simple, default link format that will pass authority and help the target page rank better. The text “Click Here” will be the highlighted link. A nofollow link looks like this:

<a href=”” rel=”nofollow”>Click Here</a>

The nofollow link will not pass any authority signals and will not directly help the target page rank better. Despite the name, however, Google will follow nofollow links — Googlebot often crawls these links, but makes a note that they don’t count for ranking purposes.

Many webmasters and SEOs use this attribute incorrectly. As we learned in PageRank, the PageRank that flows out from a page is divided by the number of links. Some SEOs still use nofollow to try to concentrate more PageRank into the other links: this does not work. Some SEOs think that Google won’t follow a nofollow link: this is not true. Google does follow them.

There are basically only three kinds of links you will want to make nofollow:

  1. Any kind of paid link, or link you were compensated for in any way, should be nofollowed. It’s against Google rules to pay for a link that passes PageRank, and a site that buys or sells links can get severely penalized by Google. Thus all ads should be nofollowed.
  2. User Generated Content (UGC) links should be nofollowed. This means anyplace where a user can write something on your page — like forums, comments, reviews — any link they include should automatically be nofollowed. You have no control over where they might link and you want to distance your site from those links, particularly if they link to spammy sites. Further, keeping them nofollow will somewhat reduce the amount of spam you get (from all the SEOs trying to use your forums/comments to build links to their own sites).
  3. Any links to sites you really, really don’t want to help. If for some reason you are linking out to a site that you are competing with for rankings (perhaps you sell information on Black Mold and want to link to the Wikipedia page but don’t want to help it outrank you) you will want to nofollow that link.

For the most case, other than ads and UGC, none of your links should be nofollow. Links to your privacy terms or to your social media account do not need to be nofollow, and it doesn’t help you if they are.

The Key to Understanding Nofollow Links

The core principle to remember with nofollow links is that making a link nofollow does not help your site in any way. All it does is prevent that link from helping another site. But your site will flow the same amount of PageRank and be crawled just the same regardless of what you do with nofollow.


Most sites have some form of pagination in them. Blogs show the most recent articles, then you can click through pagination for lists of older articles. The product listing pages of ecommerce stores have pagination to go through the entire list of products in any category.

Google used to have problems with pagination and sometimes it would rank, for example, page 3 instead of the first page. This is bad because you usually have the first page of the series optimized to be the best user experience.

Google created rel=prev and rel=next tags to better understand pagination. With these tags Google understands when its seeing a paginated series and it knows what the first page of the series is.

For very large sites, pagination can be somewhat important: for small sites it’s usually not important at all unless you know Google is ranking the wrong page of a paginated series.

Here’s how you can implement pagination:

In the <head> section of the page code, include the following for each paginated page:

<link rel=”next” href=”” />
<link rel=”prev” href=”” />

The “next” URL should be the next page in the series: so if you’re currently on page 2, the “next” will be the URL for page 3. Similarly the “prev” URL should be the previous page in the series: so if you’re currently on page 2, the “prev” will be the URL for page 1.

If you’re on the first page, you don’t need to include the “prev” line, and if you’re on the last page you don’t need to include the “next” line.

There is alternate markup that you can use if you have a View All page in addition to your Pagination. I generally don’t recommend using this, because if you have a View All page, that is the one that Google will want to return in search results, and the View All page is usually not the best user experience (if it was, you’d just use that and not have any paginated results, after all).

If you really want to have both pagination and a View All and serve the View All in the search results, instructions for that can be found here.

Duplicate Content & Rel=Canonical

Another way to control the authority flow within your site that is specific to duplicate content is rel=canonical. This code is used when your site has multiple pages that are nearly identical. This happens in almost every site of any size and isn’t necessarily a bad thing. Some site features that create duplicate content include:

  • On ecommerce sites anything that lets you change the order in which products are displayed usually creates duplicate content. If you let the user sort by best-selling, or by price (and you should) that will usually create a variant URL, but the content of the page is identical — just in a different order
  • On blogs your archive pages usually create duplicate content. For example when you look under categories, or tags, you’re getting a list of the same blog posts that exist elsewhere.
  • Ecommerce sites that use the category structure in the URL create duplicate product pages when there are different paths to navigate to a product. You might access a product both at and for example

Having internal duplicate content isn’t necessarily a bad thing. Google has even said that 25% of all the pages on the internet are duplicate pages; it’s a part of site design that makes navigation better for users and Google has learned how to deal with it.

The problem for SEOs is controlling which page ranks, and the greater issue of dilution of authority.

When Google sees two or more pages on your site that are mostly duplicates of each other (they can be slightly different and still be duplicate content) Google will choose just one of those pages to rank, and that might not be the page you want it to be.

Furthermore, perhaps some people link to one URL of your content and other people link to the other URL. Maybe you have 20 links to this great page you’ve created, but there are only 10 links to each version. Now that page only has half the links it should and isn’t ranking nearly as highly.

Enter Rel=Canonical

This is where rel=canonical comes in. This tag sits in the <head> portion of your site code and tells Google which version of a page is the canonical version — which is the official version that Google should rank. Every version should have the rel=canonical tag and they should all point to the same official URL. Then any links to any version of the page counts as if it was going to the canonical version (technically you still lose 15% of the link juice, so 85% of the authority passes).

Here’s what rel=canonical should look like, somewhere between the <head> and </head> tags:

<link rel=”canonical” href=”” />

As a best practice, every page of your site that you let Google index should have the rel=canonical tag pointing to the official version. The reason is there are a lot of ways for people to link to you that changes the URL: you could get links with a refid or UTM code (which are parameters that are used for tracking purposes) and most large sites have multiple ways to render a URL and even the engineers who work on the site code aren’t familiar with all of them.

What you have to be careful to avoid, however, is having duplicate content that each canonical to different sources. This won’t hurt you or penalize you, just Google will then decide it needs to ignore the canonical and once again makes its own decision about what to rank.

You can also point the canonical tag to an entirely different site (or subdomain). If you have two different websites, but have certain content that is identical on both sites (duplicate articles, guides, products, etc.) you can choose which site should be the canonical version.

In most cases, canonical implementation is something that you’ll need to talk to your webmaster about. They can either code them to dynamically generate based on internal logic, or they can code them so that you can set the canonical version on each page of your CMS yourself. If you have a WordPress site, you don’t need to worry about canonical tags, because WordPress takes care of that for you.

Rel=canonical is a very useful tool for making sure the correct page is ranking, and ensuring that it ranks as best as it can and it’s an important SEO best practice that should be put in place when your site is created — and as soon as possible if your site was built without it.

Dead Pages and 301 Redirects

Sometimes you have to take pages off your site. Perhaps a product or service is being discontinued, perhaps it’s outdated information that no longer belongs, or perhaps the page was just a bad idea in the first place.

Discontinuing pages is perfectly normal and happens with every site.

Where we run into trouble is when there were links from other sites pointing at those pages. Now when Googlebot crawls those links it encounters a 404 error rather than the page it was expecting to find. When that happens you do not get credit for that link — after all, the page the site was recommending isn’t there anymore, so the recommendation goes away.

Over time, discontinuing pages can having a cumulatively negative impact on SEO. You are in effect throwing out links, which are by far the most valuable and hard to gain ranking factor. Over time you’re trying to build your SEO on a pile of bones.

Happily, there’s a solution: the 301 redirect.

301 Redirects

A 301 redirect is a permanent redirect that happens at the server level (not on the HTML code of your site). When someone tries to go to the dead page, instead of getting a 404 Not Found error, then instead automatically switch to a different page (on a different URL) and load that instead.

PageRank flows through a 301 redirect, moving from the dead page to the new one.

Any way of redirecting a page that is not a 301 redirect will not pass full SEO benefit. This includes 302 (temporary) redirects and meta refresh code in the head of a page. (Some Googlers have made statements suggesting that 302s are as good as 301s: tests have proven that this is not true).

For the most part, you will need to talk to your webmaster about setting up 301 redirects. If you are using WordPress there are plugins that will let you implement them pretty easily, and odds are your hosting service probably has some kind of tool to make redirects easy to manage.

If you are using an apache server (super common for small sites) you can manually create 301 redirects by editing the .htaccess file in your root directory (this is often hidden, so you will have to set your FTP program to display hidden files — note the period in front of the filename).

Here is the code you need to program a 301 redirect directly into .htaccess:

First enable the Apache ReWriteEngine in the mod_rewirte module and enable symantic link matching by pasting this code into your .htaccess:

Options +FollowSymLinks
RewriteEngine on<

This only has to be done once, and you can then enter all the redirects you want thereafter. To actually redirect your pages use this format:

Redirect 301 /oldpage.html

There are a lot of sophisticated ways to enable 301 redirects based on regular expression (RegEx) matching, but that’s a bit beyond the scope of learning SEO. This will enable you to redirect pages as needed, and if you are moving entire sites or need specific type of page matching you should talk to your software engineer.

When to Use 301 Redirects & Where to Redirect

As a general SEO best practice, any time you take down a page, you should 301 redirect that page’s URL to another page on the site. There are certainly times when it’s appropriate to deliver a 404 error; however, if there are any links to the page you’re taking down, you really want to 301 redirect it instead and preserve that authority for your site.

As for where to redirect a page: if you’re taking a page down and there is another page on the site that provides the same (or nearly the same) info, you should redirect there. This is the easy case where a product is discontinued when the newer version of the product becomes available.

When there isn’t a nice one-for-one replacement, you generally want to redirect to the page that is one level higher in the site hierarchy. For ecommerce that means redirecting product pages to the associated sub-category or category page. Sometimes it will even mean redirecting up to the home page — but if there are links pointing to that dead page you do not want to lose them.

How to Optimize PageRank / Link Juice

Just like PageRank flow through external links, from one site to another, so too does it flow through internal links within your site. I have seen truly astonishing ranking gains simply by optimizing this internal PageRank flow — though of course you have to have good external links first!

The key factor to optimizing your internal PageRank flow is to reduce the number of links on your page as much as possible, while at the same time making sure you link to your most important pages from your global navigation. By having those links in your global navigation, you’re ensuring that you’re flowing PageRank from every single page of your site to the pages that you want to rank.

This is one of the reasons that you want to have a hierarchical site structure: your most important pages (the ones you want to rank for the most important keywords) are in the top tier in your global navigation; the next most important go a tier down and are linked to from those top tier pages, etc.

There is no golden number of links that you should have on any given page. You want only as many as you need, and no more. Most sites struggle to have fewer than 100 links and something near that 200 links is about right for a lot of sites; most sites with 500 or more links per page definitely have far too many.

Sources of Link Bloat

One of the most common culprits for link bloat is the footer. This is the least visited section of your site, and it tends to accumulate all kinds of junk links that don’t really need to be there. Sure, your site might need to have an About Us and a section about your awards, and your founders, and job openings — but those do not all need to be separate links on every single page of your site. Consider instead having the same list of text but making it all a single link that leads to an About Us page that then links out to each individual section.

Global navigation, usually the navigation up at the top of the site with dropdowns, is another common culprit. Your global navigation should link to most important keyword pages (which should usually be the pages most people are looking for anyway). In a broader sense, however, your global navigation should be used for navigation. I know that sounds obvious, but this is actually a common problem area.

Navigation vs Discovery

A typical mistake that even experienced site merchandisers do is try to boost flagging pages or products by linking to them in the global nav. The hope is that more people will see the link, and then go visit the page. What they are doing is using the global navigation for discovery instead of navigation. They aren’t trying to help someone get where they want to go: instead they’re trying to interrupt someone’s trip and convince them to go somewhere else.

I once worked on one of the largest ecommerce sites on the internet that had this issue. Over time various category managers tried to boost their categories by inserting more and more links into the global navigation drop-down menus. Many of these links appeared several different places in the global navigation.

When the Taxonomist at the company finally did a comprehensive review, he found that many of these links were generating only 50 clicks per month — this in a site with millions of visits a day. And one of those 50 clicks per month links was in the global navigation five different times, while links attracting hundreds of thousands or millions of clicks were there only once.

Navigation is not the place for discovery. If left unchecked, it makes your site increasingly difficult to navigate, harming your user experience… and of course also hurting your SEO, because now your PageRank is divided by all those useless links.

At Wayfair we once removed over 150 links from the dop-down navigation as part of an authority flow optimization. We not only saw big improvements in SEO traffic, but an A/B test showed that the new navigation improved conversion rate, and reduced usage of the internal search. Showing fewer navigation options actually made it easier to navigate the site!

There’s nothing wrong with incorporating discovery into your site, but it’s usually better done through ad blocks (where you put your own attractive ads) and it should almost never be done as a global navigation item. This can also help you to really think about whether the ad needs to be on every page of the entire site, or if it’s more appropriate to only certain sections.

SEO vs Usability

Some of the hardest decisions in site navigation is the balance between SEO and usability. Ultimately you are going to have to link to pages with zero SEO benefit: you need a link to your Contact page; you need to link to your shopping cart or RFQ form, etc. But try to be disciplined about restricting links to pages without SEO value to pages that you really do need.

In the end, usability should usually win over SEO. After all, one of the key tenants is Do You Deserve to Rank? You need a great site, and that means a site designed for people, not only for Google. But good usability is like a good break-up: you got to be brutally honest about which links you really need in there.

Which Links to Obsess About

As a general rule of thumb, you should not worry about links in the actual content of your pages. Put links there when they’re useful to users and don’t worry too much about authority flow. Unless you have a truly insane number of links regularly appearing in your content — for example if you have inane rules like “link to at least five pages from the content” — focus your attention elsewhere.

You should be obsessive about two areas of links:

  • Global Navigation Links: anything in your header, footer, and left/right nav. WordPress sites will find common culprits in tag clouds, archives, theme links, credit links (also a good way for hackers to identify your site) etc.
  • Template Level Links: If you have page types that have their own sub-navigation (like facets on product listing pages, or sub-menus)

Not only should you obsess about every single link appearing in these areas (keeping in mind: user first), but you should revisit them at least every six months. I’ve found that optimizing authority flow is like a game of whack-a-mole: you knock ’em down, but pretty soon a whole bunch more links are popping up.

A Link Calculator Tool

How many links do you have on your pages? Probably a lot more than you really think. You can use the handy Link Calculator tool to quickly get a count, and list, of every link on any page. This tool was built by my cunning engineering team at Wayfair and given to the public to use.

Understanding PageRank

Central to understanding how to flow authority through your site is an understanding of PageRank. Named for Google founder Larry Page, PageRank was the central premise that Google was designed around. In the years since Google launched it has grown in sophistication and now PageRank is just one small piece of the ranking puzzle. In terms of on-site authority work the PageRank model remains one of the most important pieces — it appears that other link-related benefits flow through a site similarly to PageRank.

PageRank is a measure of the raw link juice that flows through links into any given page on the internet.

In Google’s model, every link a page receives is like a vote of confidence saying that the page has good and useful content. The more links a page has, the more confident Google is in ranking that page.

PageRank is a measure on a scale of 1-10 of how much raw link authority a page has. But PageRank isn’t just a measure of the number of incoming links: a link from a high authority page is worth much more than a link from a low authority page. So PageRank calculates link authority based on all the incoming links to a page, and the incoming links to the pages that link to that page, and of course the incoming links to the page that links to the page that links to that page all the way back forever — it’s turtles all the way down.

The very clever thing about PageRank was the math that was able to calculate the PageRank value of every site based on an infinite chain of linking sites. But for our purposes, the important thing is that more authoritative sites are worth more.

How Do I Find Out My PageRank Is?

Put simply, you can’t.

Google used to share a version of PageRank publicly, called Toolbar PageRank. It was never a very precise measure, but it let SEOs at least track major growth in authority. Google no longer provides this, however.

Happily, you don’t need to know what your PageRank is to optimize your site.

How PageRank Flows

The PageRank of a page flows through the links on that page — or at least most of it does. Due to the way the PageRank model works, only 85% of a page’s PageRank flows out through links.

That PageRank is then divided by the number of links on a page, and a proportional amount flows through each link.

Thus if we had a page with a PageRank 10, and we had a total of 10 links on the page: 8.5 total PageRank would flow through the links (because only 85% of it flows out) and each page we’re linking to would get .85 PageRank from us. If instead we had 100 links on the page then each link would get .085 PageRank from us.

It’s worth clarifying that none of this reduces the PageRank of our page at all. The number of links you have on a page does not affect that page’s PageRank at all, and certainly doesn’t lower it. PageRank is often pictured as a bucket with water flowing out of each hole: this is the one area where that comparison breaks down. While PageRank flows through every link, it never actually takes away from the existing page. PageRank is calculated solely on incoming links from other pages.

It’s also worth noting that nofollow links (which we’ll discuss later in this section) don’t flow PageRank, but they still count as part of the link total that that PageRank gets divided by.

Thus if we had our PageRank 10 site with 5 nofollow links and 5 normal links: 8.5 total PageRank would be available to flow through the links (because only 85% of it flows out); the nofollow links would get no PageRank and the normal links would get .85 PageRank each — because we’re still dividing the available 8.5 by 10 links, even though half the links aren’t getting anything.

Almost as soon as the nofollow tag was created by Google in 2005, SEOs started “PageRank sculpting” by putting nofollow on nearly all of their internal links to control exactly how the PageRank flowed through their sites (and suffice to say they nofollowed every external link). Google very quickly (also in 2005) changed the PageRank calculation so that nofollow links still counted when calculating the PageRank that flowed through links (as I describe above).

The only way you can increase your PageRank is by getting more links pointing to the page — and preferably links from higher PageRank sites. The only way you can control how much PageRank flow out through your links is by changing the number of total links on the page.

PageRank vs Link Juice

Technically PageRank is one part of the authority that passes through links that Google measures. SEOs often use the term Link Juice to more broadly describe all benefits that pass through link; however, in the SEO world PageRank and Link Juice are used interchangeably.

When Google talks, however, and they talk about PageRank, they usually mean just PageRank, and not necessarily all authority factors that pass through links.


On-Site Authority Intro

An important part of building authority is managing your site structure to make optimal use of the authority that your site earns through incoming links. In the SEO world this is often referred to as technical SEO, since it involves some of the more technical code work done in SEO.

A colleague of mine often referred to this as “SEO Black Magic” because it’s the part where you do things invisible to users that can, in certain situations, give magical-seeming improvements to rankings (this is rare).

Ultimately your authority is going to come from other sites linking to yours, and the way you do that looks a lot more like traditional marketing. The point of on-site authority work is to ensure that you get the most benefit out of every incoming link. A poorly optimized site bleeds authority, essentially throwing it out the window rather than taking advantage of it to boost rankings.

A well optimized site makes the very most out of every drop of link juice coming in.