Website Helpers.com

  Articles, tips, and resources for webmasters

a project by Michael Bluejay | email

Inbound links increase ranking

In general, the more links to a page, the higher the page will rank in the SERPs. We'll have a lot more to say about this below, but for now just understand that every link to a page is a "vote" for that page.

I have mixed feelings about including this section because I know readers will jump all over the details and think of them as "tricks", and come up with a million related questions about whether such-and-such linking strategy is good or bad, and how it compares to some other linking strategy. To do so is to miss the point. First of all, what's below is all you need to know about links for search ranking purposes. Second, the reason you need to know only this and no more is that you should be focusing on building a quality site, not obsessing over links. If you build a great site, other webmasters will link to it without your even asking them to.

But let's proceed anyway.

Anchor Text

The anchor text is the words that appear in a link. For this reason it's also called the link text. The engines pay special attention to the anchor text, because it seems reasonable that if they finds a link to, say, how to buy a house, then the page being linked to is probably about how to buy a house.

Once I tested the power of anchor text by linking to a page with the anchor text "pancreas of fury". Soon the page I linked to was #1 in Google for a search on "pancreas of fury", even though the words "pancreas" and "fury" didn't appear anywhere on the page at all! It even outranked the page that contained the anchor text itself. (Don't bother searching for this page; I took it down long ago.)

The most famous exploit of anchor text was when bloggers teamed up to all link to George W. Bush's page using the anchor text "miserable failure". That's why when you search Google for "miserable failure" Bush's page comes up first. (Republicans retaliated by linking to Jimmy Carter's and Michael Moore's pages with that phrase to get those sites to show up as well.)

One way of trying to use anchor text to your advantage is with your navigation links. For example, instead of the link text "Home", you might use something like "Baby Bracelets Home", if your site sells baby bracelets. When other webmasters link to you, you can also see if you can get them to include your keywords as the anchor text.

Be warned that some consider anchor text to be an ingredient of the theoretical overoptimization penalty mentioned above. The idea is that having a lot of backlinks with the same anchor text that's in the <TITLE> tag and all over the page triggers the penalty. While I don't know for sure whether this is the case, my feeling is that there is probably such a penalty but it's not a simple black and white issue, and it would be difficult to determine under exactly what circumstances it's triggered.

Links in the body copy

It is believed that links that are in body copy (inside a paragraph) count for more than links that are all by themselves in a navigation bar, sidebar, or footer. As for clickability, there is little question that links inside body copy get clicked much more than links at the edges of the window. A fellow webmaster recommends my site by putting a link to mine in the footer of his site, on all 600 page. Yet most of my traffic from his site comes from a link within an article where he mentions my site. Content links are more likely to be clicked, and many believe that they count for more in the engines, too.

Internal Links

The general feeling is that you get less credit for links from within your own site rather than links that come from external sites, but this doesn't mean you should neglect internal links. After all, internal links are the easiest to get. The first thing is to make sure you link to as many of your pages as possible from your home page. If you can't fit more links in your navigation menu, put some extra links on the bottom of the page. The idea is that you want as few "hops" as possible to get to all your pages. If a user (or Googlebot) has to go from your Home Page > Category Page > Article Page, that's two hops, or "two-deep". PR is being passed down the chain, and gets weaker as it wends its way down. The Home Page gives a little PR to the Category Page, and then the Category Page gives a little of what IT has to the Article Page. If you have fewer hops then the pages get more PR passed to them.

Some people confuse hops with directories. They think that <domain.com/subdir/subdir/file.htm> is bad because it's too far from the home page. But it doesn't matter how far down the file tree a file is, what matters is how many clicks it takes to get to it. If the home page links directly to the page mentioned above then it's just one hop, which is as good as you can get.

Each and every page in your site should link back to the home page, as well as to a handful of the most important pages on the site, and also to pages related to the page in question.

Larger sites (those with a larger number of pages) seem to do better in the engines, and this probably has something to do with all the links within the site. Though Brett Tabke's "Build a Successful Site in Twelve Months" is somewhat dated, his advice to build one new page of content per day is as sound as ever. But even if you can't create content at anything close to that rate, just build what you can. If you can't build one a day then try one a week. If you can't build one a week try one a month. The point is, the more the better.

If you have navigation links in Javascript or Flash, make sure you also have links in plain HTML so the engines can see those links.

PageRank

PageRank (PR) -- Google's measurement of how important a page is.

PageRank ("PR") is Google's measurement of how "important" a page is, based on how many pages link to that page, and how important those pages are themselves. Pages are ranked on a scale of 0 to 10, with 10 being "extremely important". The name PageRank is ironic: it's actually named after Google founder Larry Page who came up with the idea, not named after web pages.

You can see the PageRank of any page by installing the Google Toolbar, though the PageRank information displayed isn't necessarily accurate, which is just one reason you should generally ignore your PR score, as I'll explain below.

Other engines use a similar scheme to evaluate incoming links but they call it something else. We'll use the term PageRank in this discussion even though we mean to talk about ranking on all the engines as a whole, not just Google.

PageRank is just one of the factors an engine uses to rank pages in the SERPs, and it's not the most important. A page's relevance to what's being searched for is way more important than PageRank. It's definitely not the case that the page with the highest PR will always rank first. A PR4 page can easily be listed above a PR6 page if the engine figures that the PR4 page is more relevant to the query being searched. All things being equal, the page with the higher PR will rank higher. But things are never equal.

Pages get PR by being linked to from other pages. More incoming links means a higher PR. And the higher the PR of the linking page, the higher the PR of the page being linked to. Google has published the formula for computing PR but it's very complex, and is irrelevant to the typical webmaster besides. (Remember to focus on the general things, not the specific things.) In general, a page will have a PR of at least one less than the highest-PR page linking to it. For example, if a PR7 page links to your home page, your home page will probably be at least PR6.

The more outbound links on a page, the less PR will be transferred on each outgoing link. On a page with two links, each link transfers out half the PR as a page with one link. Some webmasters avoid linking to external sites because they want to keep all their PR for themselves, for their internal links. Some of them even put external links in Javascript because PR is only passed through HTML links. This is probably a mistake. While it's true that such methods may preserve PR, we have to remember that PR isn't the only factor in ranking well. Another factor is whether your site appears credible andwilling to share information, and your site appears more credible and more willing to share information when you link out to quality sites. My guideline for myself is simple: I link out where appropriate, without considering PR.

While you can dilute amount the PR transferred by increasing the number of links on a page, this never affects your PR score. If your page is PR5, it will remain PR5 no matter how many links you put on it. Adding links simply dilutes each link's power, but does not change the page's PR score.

Webmasters actively seek links to their sites from other sites because that increases their PR and their rankings, in theory. You can certainly solicit links but I think your time is better spent making your site a natural link magnet. And in any event, I suggest you ignore your numeric PR score, as I'll explain below.

You don't get PR immediately. It can take a while, sometimes several months, though most sites get their initial PR within three months. Though again, you shouldn't really care what your PR is.

Don't become desperate for incoming links. Mike Grehan tells the story of how some webmasters were so insensitive they put their urls into the guestbook on a memorial page for Grehan's deceased friend -- just so they could have more links coming to their sites. That's really sad.

I suggest you ignore PR. Focus on making your site bigger and better, and don't worry about what your PR is. Jill Whalen of HighRankings.com says it best:

1. The PR you see on your Google toolbar is only a representation of your page's true PageRank, which only Google really knows.

2. The Google toolbar is flaky. As you've seen, it can show one number at one time and another at another time. This has nothing to do with YOUR page; it's simply a toolbar issue.

3. Google's PageRank number as shown by the toolbar means next to nothing. It's not something you need to chase after, look at, worry about, wonder about, or care about....

My advice for you would be to turn off the green bar aspect of your toolbar and move on with your life. Nothing lost, nothing gained. You'll thank me for it later.

If you really want to know more about PR, I suggest Danny Sullivan's guide to PageRank.

Backlinks

A backlink is just a fancy name for a link to your page from another domain. Webmasters want more backlinks because more backlinks mean better ranking in the SERPs. You can see some of the pages link to your domain by searching Google for site:example.com (substituting your own domain name, of course). Google usually doesn't show all the pages that link to yours, just a smaller subset. To see the links to a particular page, search for site:http://example.com/page.html

The number of backlinks you have is referred to as link popularity.

If you create a high quality site then other webmasters will link to your site -- or to useful internal pages -- without your asking, increasing your link popularity.

Linking to a page within a site (as opposed to the home page) is called deep linking, and there's nothing wrong with it. In fact, it's better to link to an inside page if that's the information you're trying to point a visitor towards. (See more about how Google ranks pages, not sites.)

Reciprocation

Since webmasters want more links, many of them try to trade links with other sites, offering to link to another site in exchange for a return link. Such links are called reciprocal links. There is disagreement over whether the engines discount reciprocal links because they're obvious trades. It's even possible that link trades can hurt your ranking, if you trade with a promiscuous site that links to anybody. (More on this in a minute.) My hunch is that the engines do count reciprocal links, but just don't value them as highly as one-way links.

Personally, I feel that whether someone links to you is a poor criteria for whether you should link to them. I will link to another site if I think it's high quality and will be of interest to my readers -- regardless of whether the other site links back to me or not. And when I link to another site I generally put the link at the bottom of an appropriate content page, rather than on a separate Links page.

Since backlinks are so desirable, webmasters spend a lot of time emailing other webmasters to request reciprocal links. The overwhelming majority of these requests are poorly thought out and don't get a favorable reply. Webmasters who send out form letters begging for links are known by the unflattering name link whores. Here's my separate article on how NOT to request reciprocal links.

Rather than writing to others to ask that they trade links, I instead link to sites I find valuable without asking for a return link. I try to encourage links to my own sites by building high-quality sites that other webmasters will naturally want to link to. For the most part I restrict link-begging to my own site, not to other webmasters' mailboxes -- see the humorous (I hope) pink ad in the sidebar at left, "We'll cry if you don't link to us." Part of this is a parody of the whole link begging phonemenon that webmasters seem to be obsessed with.

I finally got fed up with all the link trade requests I received every day that I put a filter in my email to trash those requests before I even see them. That's how valuable I think link trades are.

Whether you're trading links or not, put your outbound links on your content pages where possible, not on a separate links page. If you do have a links page, make it into a useful directory by organizing the various links into categories, and giving a short one-paragraph description of each site you link to. A good links page stands on its own as a useful resource, and isn't just a dumping ground for reciprocated links.

Link Farms and Directories

Link Farm -- A large, worthless collection of unorganized links, with little to no descriptions, and no respect to quality.

Some webmasters will trade links with anybody in an effort to get backlinks. This is a classic case of buying the low-fat cookies instead of eating your vegetables. They fill up pages with dozens or hundreds of unorganized links with no description. Most people would consider such pages worthless, and the engines feel the same way. They consider them to be link farms, and when the engines identify a link farm they turn off its ability to pass PR through its outbound links. Sometimes this penalty takes the form of the site getting its PR set to zero by Google as shown in the Google toolbar, or the page might still be set to no longer pass PR to other sites even though the toolbar shows the site to still have PR. (More below on sites not passing PR.)

The engines consider link farms to be bad neighborhoods, and if you associate with them that bad reputation can rub off on you. Your site can get penalized for linking to a bad neighborhood. This is yet another reason why you should link only to high quality sites.

Some webmasters have an unfounded fear about being linked to from a link farm. Don't worry, you'll never be penalized because of who links to you, because that's beyond your control. But you can control who you link to, so it's only when you link out to a bad neighborhood that you can get into trouble. Don't link to a site that you have a bad feeling about, seems low quality, or which is PR0. (PR0 could be the result of a penalty. It could also be because the site is new, but there's no easy way to tell the difference conclusively.)

On the other side of the coin are directories. The engines like directories because they organize information, and Google's mission is to make information easily accessible. [Here I used to link to three example sites I managed, which were at or near the top of the SERPs for the things they covered, but I no longer run those sites.  While good directories rank well, they're a lot of work, which is why I don't do those sites any more.]

At first glance link farms and directories have a lot in common -- they're both collections of links to other sites. So what's the difference? I would say the difference is usefulness. The entries in directories are organized by topic and have enough description (or are organized well enough) that you have an idea about what the sites are about before you visit them. A link farm is usually just a bunch of links slapped on a page with no useful order or descriptions.

How do the engines tell them apart? I don't know how their algorithm works, but somehow they seem to do a pretty job job of making the distinction, since you're more likely to see directories in the top of the SERPs instead of link farms. But it stands to reason the the engines sometimes improperly tag directories as link farms and vice-versa.

While it's easy to identify the best directories as directories and the worst link farms as link farms, obviously there is some gray area for sites in the middle. What if a list of sites isn't so hot but isn't completely useless? The engines will make their own judgements, but as for you I suggest that you don't link to a site unless you consider it high quality, and if you build your own directory, make it high quality as well.

Links from legitimate directories are invaluable. Find directories that cover your industry and ask to be included in them. Search Engine Watch members have compiled a list of directories they like.

Buying and Selling Links

Because links are so important some webmasters buy links to their sites from other sites, but Google frowns on this wholesale marketing of links. The obvious question that comes up is, what's the difference between a sold link and a text ad? The main difference is probably intent. If it's sold for the purpose of driving traffic to the advertiser's site then it's an ad. If it's sold for the purpose of increasing the advertiser's PageRank then it's a paid link. So how do we discern the intent? Probably by how it's marketed. If it's marketed as a paragraph ad, with the selling feature being how much traffic the publisher's site gets, and no mention is made of PR (either the publisher's or the advertiser's), then it seems like an ad. If it's marketed as a "link", with the selling feature being how much PR the publisher's site has, then it seems like a paid link.

How can the search engines tell the difference? I don't know, and I don't worry about it. As always, I focus on the general things and not the specific. I try to do what Google wants and to focus on site quality.

Should you buy links to your site? No. Your site should be good enough that other webmasters will link to it without your paying for it. If it's not good enough, then you need to spend time improving your site rather than trying to trick your way to the top.

Sample Text Ad Website Helpers has useful articles for both beginning and experienced webmasters, and it's free. WebsiteHelpers.com

You can certainly buy text ads (paragraph ads which contain a link), because that's advertising and you're hoping that some percentage of readers will click over to your site. The difference is that buying links is trying to game the system, while advertising is a legitimate, traditional practice. Yes, you will get some link benefit from a text ad since the text ad contains a link, but you should think of that as a side effect/bonus of the advertising that you're buying for advertising purposes.

Should you sell links from your site? I advise against it. Besides the fact that you risk a search engine penalty, it's just kind of cheap. Again, you can certainly sell full-blown text ads which contain a link. I do that all the time.

Incidentally, the text ads that are displayed in Google Adwords/Adsense program don't count towards backlinks, because the ads are delivered with JavaScript and the search engines don't index the JavaScript output on a page.

Pages not passing PR

If an engine feels that a page is a link farm or that it's selling links, the engine can prevent the page from passing PR. The sites being linked to will not get a PR boost.

Should this matter to you? Probably not. As I keep saying, I suggest you ignore PR. But let's assume you're determined to find out whether a page is passing PR, because you're either buying or selling links to or from that page (even though I recommend against that, too).

So how do you tell whether a page is passing PR? There's no quick and easy way, but there's a cumbersome way: Create a brand-new page and link to it from the page you suspect is not passing PR. The page you link to should generally get a PR of one less than the source page. (If the source page was PR5, the new page should be PR4. Use the Google Toolbar to check the PR.) If the new page doesn't get any PR within three months, it's a good bet that the page isn't passing PR. Note that it usually doesn't take three months for Google to pass PR if it's going to, but sometimes it can. In any event, if two months have passed on the new page doesn't have PR yet, there's no way to tell whether the lack of PR is because Google isn't passing PR from the source page, or whether Google simply hasn't gotten around to passing the PR.

Link Age

There is some agreement that older links count for more than recent links. Anyone can start a new site and buy thousands of links to it. The easiest way the engines can keep webmasters from screwing with the SERPs in that manner is to discount the value of new links. The theory is that the older a link is, the more weight it carries.

Relevance and Authority

Links from relevant pages count for more than links from unrelated pages. If a page ranks well for what you want to rank for, then the engine considers it relevant. The engines determine relevance by considering such things like the <TITLE> tag, body copy, and the nature of the sites linking in and out of the page in question. This just scratches the surface on this topic; here's a Webmaster World thread on relevancy. Though as one poster there says, focusing on the nuts and bolts of how relevancy is calculated is probably missing the point.

A related concept is authority pages. An authority page is one that has a lot of inbound links itself for a particular topic. If a page seems to be an authority on, say, Perl programming, then its outbound links to external pages about Perl carry more weight than links from a non-authority page.

Suspicious Activity

While we don't know exactly which criteria the engines use to determine the legitimacy and importance of links, we do know that it's not hard for them to evaluate links however they want. For example, if a brand-new site suddenly appears out of nowhere with thousands of pages of content, has thousands of links pointing to it, and most of them use the same exact link text, and most of them are on every page of a site, by themselves, separated from the content, then they may look like paid links and not natural links, and may not count for much. It's unlikely that such a site would actually be penalized though (having its rank moved down), because if that were the case it would be easy for a webmaster to penalize a competitor by buying a bunch of cheap links to the competitor's site. It's important to distinguish between a method not helping vs. the method invoking a penalty.

Splitting PR

Splitting your PR -- Wasting credit for inbound links because some links go to the www. version and some don't.

Conventional wisdom has it that if some sites link to your page with the www and some don't, you'll get credit only for the links to the version that shows up in the SERPs. That is, if Google lists your site as domain.com, then you'll get credit only for the links to domain.com, and not to www.domain.com. And if Google lists your site as www.domain.com then you get credit only for the links to www.domain.com and not to domain.com. Or so the thinking goes. This wasted credit is often called splitting your PR.

Since you can't control how people will link to you, the preferred solution is to have your server either automatically add the www. if the link didn't contain it, or to strip the www. if it did. When I decided to try to avoid splitting my PR on some of my sites, I chose to have the www. stripped, because I've never liked the www. When someone follows a link to www.michaelbluejay.com, the server automatically strips the www. and so the user then sees just michaelbluejay.com in the address bar.

A warning: Yahoo penalized one of my sites into oblivion until recently, and the most likely reason was that I had the server remove the www. I fixed the problem by forcing the www, and then when Yahoo indexed my site I switched back to the non-www. I haven't had a problem removing the www. on any other sites, but if you remove the www. and then fall out of favor with an engine, try doing the opposite, and forcing the www. instead. My problems with Yahoo are detailed under non-www penalties.

Here's the code to remove the www. Put this in the < .htaccess> file, described under redirecting pages.

RewriteEngine on 
 RewriteCond %{HTTP_HOST} !^yourdomain\.com 
 RewriteRule (.*) http://yourdomain.com/$1 [R=permanent,L]

And here's the code to force the www:

RewriteEngine on
 RewriteCond %{HTTP_HOST} ^yourdomain\.com
 RewriteRule (.*) http://www.yourdomain.com/$1 [R=permanent,L]

Summary of Link Factors

Some of these factors are assumed rather than being known definitively, as described above.

Carries more weight	Carries less weight
From higher PR page From another domain Within the body copy Keywords in link text From a relevant page (page contains keywords, esp. in page TITLE) From an authority page Old	From lower PR page From the same domain By itself, not as part of a paragraph Kewords not in link text From an unrelated page From a non-authority page From a link farm New

Now continue this series below...