Chapter 15

Using Link Popularity to Boost Your Position

In This Chapter

Understanding how search engines calculate value

Building Web site hubs

Weighing pages with PageRank and TrustRank

Using keywords in links

Identifying links that aren’t links

Thousands of site owners have experienced the frustration of being unable to get search engines to index their sites. You build a Web site, you do your best to optimize it for search engines, you “register” with the search engines, and then nothing much happens. Little or no traffic turns up at your site, your pages don’t rank well in the search engines, and in some cases you can’t even find your pages in the search engines. What’s going on?

Here’s the opposite scenario: You have an existing site, one that’s been around for a while but with few links pointing to it, and finally find a few other sites to link to it. You make no changes to the pages themselves, yet all of a sudden, you notice your pages jump up in the search engines.

There’s a lot of confusion about links and their relationship to Web sites. Many site owners don’t even realize that links have a bearing on their search engine positions. Surely, all you need to do is register your page in a search engine and it will be indexed, right? Nope. This chapter takes the confusion out of links by showing you how they can help, and what you need to know to make them work.

Why Search Engines Like Links

A little over a decade or so ago, pretty much all you had to do to get your site listed in a search engine — and maybe even ranked well — was register with the search engine. Then along came Google in 1998, and all that changed. Google decided to use the links from other Web sites as another factor in determining whether the site was a good match for a search. Each link to a site was a vote for the site, and the more votes the site received, the better a site was regarded by Google.

Google: All about links

In fact, Google’s original name was BackRub. No, really. Check out Google’s own corporate history (www.google.com/corporate/history.html):

1996: Larry and Sergey, now Stanford computer science grad students, begin collaborating on a search engine called BackRub.

1997: Larry and Sergey decide that the BackRub search engine needs a new name. After some brainstorming, they go with Google.

BackRub? As in backlinks, links pointing to a Web page. And as in “you rub my back, and I’ll rub yours,” I think, if the original logo is anything to go by. You can see that in Figure 15-1, and again (no, really), this is the actual original BackRub/Google logo.

The founders of Google came up with the original name for a good reason; the original project was all about linking. The research project that led to Google was, according to Wikipedia, “focused on the problem of finding out which Web pages link to a given page, considering the number and nature of such backlinks to be valuable information about that page.”

So there you have it. Google was revolutionary because of the weight given to links. Google changed the game totally, and now all the major search engines look at links to learn about the referenced pages.

Links: The gift that keeps on giving

Links pointing to your Web pages do several things for you:

Links make it easier for search engines to find the page. As the searchbots travel around the Web, they follow links. They index a page, follow the links on that page to other pages, index those pages, follow the links on those pages, and so on. The more links to a page, the more likely the page is picked up and indexed by search engines, and the more quickly it happens.

Search engines use the number of links pointing to a page as an indication of the page’s value. If lots of pages link to your page, search engines place a greater value on your page than pages with few links pointing to them. If you have lots of links from sites that are themselves linked to by many other sites, search engines conclude that your site must really be important. (Google calls this value the page’s PageRank, but Google isn’t the only search engine to use links as an indication of value; Yahoo!, for instance, before the Yahoo!/Bing partnership, was using something called Web Rank, and Bing has Page Score.)

Links provide information to search engines about the page they’re pointing to. The link text often contains keywords that search engines can use to glean additional information about your page. The theme of the site pointing to your site may also give search engines an indication of your site’s theme. For example, if you have links from hundreds of rodent-related Web sites, and those links have rodent-related keywords in them, it’s a good bet that your site has something to do with rodents.

Links not only bring searchbots to a page, but may also bring people to the page. The whole purpose of your search engine campaign is to bring people to your site, right? Sometimes people will actually click the links and arrive at your site; people often forget that in the SEO world!

Links are very important. Sometimes they mean the difference between being indexed by a search engine and not being indexed, or between being ranked well in a search engine and barely being ranked at all. In this chapter, I delve into this subject, a topic broadly known as link popularity, to give you a good understanding of what links from other sites pointed to yours is all about. Then, in Chapters 16 and 17, you discover how to get other sites to link to yours.

Backlinks are an integral part of the optimization of your Web site. A backlink — this may surprise you — is a link back to your site. Search engines look at backlinks to figure out what your site is about and how important it is. Links aren’t something detached from your site; they’re an integral part of your site.

Think of your Web site in terms of a regional map. Your site is the major city, and backlinks are the roads bringing traffic into the city. A geographer looking at the map wouldn’t regard the city and roads as separate entities; they’re all part of the same economic and social system. So don’t think of the links pointing to your site as something “out there” — they’re a critical part of your site.

Search engines are trying to figure out what site or page is the best match for a search. As you discover later in this chapter, search engines use links as one way to determine this. As with content though (discussed in Chapter 10), using the number of links to and from a site to measure significance is an imperfect method. A page can conceivably be the best page on a particular subject, yet have few links to it. Just because I publish a page today doesn’t mean it’s worse than a page that was published five years ago and now has many links to it. However, search engines have difficulty figuring out what the searcher needs, so they have to use what information is available to them. Using links is a way of recruiting Web site owners to help point out useful sites and pages. The strategy is imperfect, but that’s the search engine world you’re living in.

Understanding Page Value and PageRank

Search engines assign a value to your site based on the links pointing to it. The most popular term for this kind of ranking is PageRank, which Google uses. The PageRank is a value that Google gives to a page, based on the number and type of links into the page.

PageRank is mentioned frequently in the search engine optimization field for several reasons:

Google is the world’s most important search engine and will remain so for the foreseeable future.

You don’t need a PhD to figure out a page’s PageRank. The Google toolbar — available for various browsers (see http://toolbar.google.com) — shows you the currently loaded page’s PageRank. (Okay, strictly speaking that’s not true, but it does provide an indication of the PageRank, which I discuss in more detail later in this chapter.) If you’re using a browser that can’t use the Google toolbar — such as the most recent version of Firefox — there are various other SEO toolbars that can show you PageRank.

I also have a basic idea of how PageRank is calculated. Sergey Brin and Lawrence Page, founders of Google, published a paper about the algorithm while at Stanford University. It’s almost certain that the algorithm used by Google now isn’t the same as the one they published, but the concepts will be similar.

Although this section focuses on PageRank, other search engines use similar rankings, and the things you do with links that boost your PageRank also help boost your site with other search engines.

PageRank — one part of the equation

The PageRank value is just one part of how Google determines which pages to show you when you search for something. I want to stress that point because so many people get really hung up on PageRank. A low PageRank is often an indicator of problems, and a high PageRank is an indicator that you’re doing something right, but PageRank itself is just a small part of how Google ranks your pages.

When you type a search term into Google and click Search, Google starts looking through its database for pages with the words you’ve typed. Then Google examines each page to decide which pages are most relevant to your search. Google considers many characteristics, such as what the <TITLE> tag says, how the keywords are treated (are they bold, italic, or in bulleted lists?), where the keywords sit on the page, and are keywords in links pointing to the page (in links both outside the site and within the site). Google also considers PageRank. Clearly, a page with a low PageRank could rank higher than one with a high PageRank in some searches. When that happens, it simply means that the value provided by the high PageRank isn’t enough to outweigh the value of all the other characteristics of the page that Google considers.

I like to think of PageRank as a tiebreaker. Imagine a situation in which you have a page that, using all other forms of measurement, ranks as well as a competitor’s page. Google has looked at both pages, found the same number of keywords in the same sorts of positions, and thinks both pages are equally good matches for a particular keyword search. However, your competitor’s page has a higher PageRank than yours. Which page ranks higher in a Google search for that keyword? Your competitor’s.

Here’s something else that comes with PageRank, by the way: The higher the PageRank, the more often the page is likely to be re-indexed.

Many people claim that site owners often focus too much on PageRank (that may be true) and that, in fact, PageRank isn’t important. But PageRank (or something similar) definitely is a factor. As Google has said:

“The heart of our software is PageRank™, a system for ranking web pages developed by our founders Larry Page and Sergey Brin at Stanford University. And while we have dozens of engineers working to improve every aspect of Google on a daily basis, PageRank continues to provide the basis for all of our web search tools.”

So, Google claims that PageRank is in use and is important. As long as it includes a PageRank indicator on the toolbar, I have to think it must still be in use, unless Google is playing a huge joke on everyone (which I wouldn’t totally discount, I must admit). But you need to keep its significance in perspective. It’s still only part of the story.

It all comes down to what the searcher is searching for. A page that ranks well for one keyword or phrase may rank poorly for another. Thus, a page with a high PageRank can rank well for some keywords and badly for others.

The PageRank algorithm

I want to quickly show you the original PageRank algorithm; but don’t worry; I’m not going to get hung up on it. In fact, you really don’t need to be able to read and follow it, as I explain in a moment. Here it is:

PR (A) = (1 – d) + d (PR (t1) / C (t1) + ... + PR (tn) / C (tn))

Where:

PR = PageRank A = Web page A d = A damping factor, usually set to 0.85 t1...tn = Pages linking to Web page A C = The number of outbound links from page tn

I could explain all this to you, honestly I could. But I don’t want to. Furthermore, I don’t have to because you don’t need to be able to read the algorithm. For instance, do you recognize this equation?

Don’t think you can kid me. I know you don’t know what this is. (Well, okay, maybe you do, but I’ll bet over 95 percent of my readers don’t.) It is the Law of Universal Gravitation, which explains how gravity works. I can’t explain this equation to you, but I really don’t care because I’ve been using gravity for some time now without the benefit of understanding the jumble of letters. The other day, for instance, while walking down the street, someone shoved a flyer into my hand. After walking on, glancing at the flyer, and realizing that I didn’t want it, I held it over a trash can, opened my hand, and used gravity to remove it from my hand and deposit it into the trash can. Simple.

Rather than take you through the PageRank algorithm step by step, here are a few key points that explain more or less how it works:

As soon as a page enters the Google index, it has an intrinsic PageRank. Admittedly, the PageRank is very small, but it’s there.

A page has a PageRank only if it’s indexed by Google. Links to your site from pages that have not yet been indexed are effectively worthless, as far as PageRank goes — actually, as far as any linking benefit goes. If it’s not in the index, it can’t confer any kind of benefit to your site.

When you place a link on a page, pointing to another page, the page with the link is voting for the page it’s pointing to. These votes are how PageRank increases. As a page gets more and more links into it, its PageRank grows.

Linking to another page doesn’t reduce the PageRank of the origin page, but it does increase the PageRank of the receiving page. It’s sort of like a company’s shareholders meeting, at which people with more shares have more votes. They don’t lose their votes when they vote; next time a meeting takes place, they have just as many votes. But the more shares they have, the more votes they can place.

Pages with no links out of them are wasting PageRank; they don’t get to vote for other pages. Because a page’s inherent PageRank is not terribly high, this isn’t normally a problem. It becomes a problem if you have a large number of links to dangling pages of this kind. Though rare, this could happen if you have a page that many external sites link to that then links directly to an area of your site that won’t benefit from PageRank, such as a complex e-commerce catalog system that Google can’t index or an external e-commerce system hosted on another site. Unless the page links to other pages inside your Web site, it won’t be voting for those pages and thus won’t be able to raise their PageRank.

A single link from a dangling page can channel that PageRank back into your site. Make sure that all your pages have at least one link back into the site. This usually isn’t a problem because most sites are built with common navigation bars and often text links at the bottom of the page. Of course, the links have to be readable by the search engines (see Chapter 8).

The page receiving the inbound link gets the greatest gain. Ideally, you want links into your most important pages — pages you want ranked in search engines. PageRank is then spread through links to other pages in the site, but these secondary pages get less boost.

It’s important to understand that Web sites don’t have a PageRank; Web pages have a PageRank. It’s possible for a site’s home page to have a high PageRank, although its internal pages rank very low.

Here are a couple of important implications from this:

You can vote large amounts of PageRank through your site with a single link. A page with a PageRank of 5 can pass that to another page as long as it doesn’t split the vote by linking to other pages.

When I use the term pass, I use it in the sense of passing on a virus, not passing a baton. Linking from page A to page B passes PageRank from A to B in the same way that person A may pass a cold to person B. Person A doesn’t get rid of the cold when he passes it to B; he’s still got it. Likewise, page A still has its PageRank when it passes PageRank on to page B.

You can ensure that PageRank is well distributed around your Web site by including lots of links. Linking every page to every other page is the most efficient way to ensure even PageRank around the site.

Measuring PageRank

How can you discover a page’s PageRank? You could install the Google toolbar into your browser; go to www.toolbar.google.com. If it’s not available for your browser, you might install a different toolbar — you can find various toolbars that provide PageRank in your browser’s add-ons library.

Each time you open a page, the toolbar will load the page’s PageRank, as shown in Figure 15-2. In the case of the Google toolbar, if the bar is all gray, the PageRank is 0. If it’s all green, the PageRank is 10. You can estimate PageRank simply by looking at the position of the green bar, or you can mouse-over the bar, and a pop-up appears with the PageRank.

If the PageRank component isn’t on your toolbar, click the Options button on the right of the bar and then select Options from the menu that appears. In the Toolbar Options dialog box, click the Privacy tab, click the Use PageRank check box, and then click the Save button.

If you don’t want to install a toolbar, you can still check PageRank. Search for the term pagerank tool to find various sites that allow you to enter a URL and get the PageRank.

It’s important to understand that as far as PageRank goes, zero is not zero, and ten is not ten. Although commonly referred to as PageRank, and even labeled as such, the PageRank value provided by PageRank tools and services is not the page’s actual PageRank. It’s simply a number indicating the approximate position of the page on the PageRank range. Therefore, pages never have a PageRank of 0, even though most pages show 0 on the toolbar, and a page with a rank of, say, 2 might actually have a PageRank of 25 or 100.

The true PageRank scale is probably a logarithmic scale. Thus, the distance between PageRank 5 and 6 is much greater than the difference between 2 and 3. The consensus of opinion among people who like to obsess over such things is that PageRank is a logarithmic scale with a base of around 5 or 6; some people believe it’s more likely base 8, or perhaps higher.

Suppose for a moment that the base is actually 5. That means that a page with a PageRank of 0 shown on the toolbar may have an actual PageRank somewhere between a fraction of 1 and just under 5. If the PageRank shown is 1, the page may have a rank between 5 and just under 25; if 2 is shown, the number may be between 25 and just under 125, and so on. A page with a rank of 9 or 10 shown on the toolbar most likely has a true PageRank in the millions. With base 5, for instance, the toolbar PageRank would represent a true PageRank as shown in Table 15-1.

Table 15-1 Pure Conjecture — What Toolbar PageRank Would Represent if PageRank Were a Logarithmic Scale Using Base 5

Toolbar PageRank	True PageRank
0	0–5
1	5–25
2	25–125
3	125–625
4	625–3,125
5	3,125–15,625
6	15,625–78,125
7	78,125–390,625
8	390,625–1,953,125
9	1,953,125–9,765,625
10	9,765,625–48,828,125

The maximum possible PageRank, and thus this scale, continually changes as Google recalculates PageRank. As pages are added to the index, the PageRank has to go up. Periodically, Google recalculates PageRank Web-wide, and PageRank drops for many sites, perhaps most.

How can you be sure that the numbers on the toolbar are not the true PageRank? The PageRank algorithm simply couldn’t work on a scale of 1 to 10 on a Web that contains billions of Web pages; it just wouldn’t make sense. It’s not logical to assume that sites like Yahoo! and Google have a PageRank just slightly above small, privately owned sites. I’ve had pages with ranks of 6 or 7, for instance, whereas the BBC Web site, one of the world’s 50 most popular Web sites according to Alexa, has a PageRank of 8 or 9. It’s not reasonable to assume that its true PageRank is just 50 percent greater than pages on one of my little sites.

Here are two important points to remember about PageRank values provided by these toolbars and services:

Two pages with the same PageRank may actually have a very different true PageRank. One may have a PageRank that is a quarter or a fifth of the other, perhaps less.

It gets progressively harder to push a page to the next PageRank on the toolbar. Getting a page to 1 or 2 is pretty easy, but to push it to 3 or 4 is much harder (though certainly possible). To push it to the higher levels is very difficult indeed — 8 or above is rare.

Leaking PageRank

It’s possible for PageRank to leak out of a site, despite the fact that pages don’t lose PageRank when they link to other pages. Here’s how: Each time you link from one page to another, the origin page is voting for the recipient page. Thus, a link from one page in your site to another page in your site is a vote for that other page. If you link to a page on another site, you’re voting for another site’s page rather than your site’s page.

Suppose that you have a page with a PageRank of 10,000, and it has 40 links on it. Each link is getting a PageRank vote of 250 (250 × 40 = 10,000). Now suppose that half the links on the page are external. In that case, you’re taking 5,000 votes and pointing to pages out of your site rather than inside your site. So PageRank leaks in the sense that the PageRank of some pages in your site is lower than it could be.

Generally, you should worry more about getting backlinks to your site from appropriate Web sites than about how much PageRank is leaking through your outgoing links. You can build PageRank quickly by using the techniques in Chapters 16 and 17, and in most cases, worrying about outgoing links won’t save you much PageRank. Still, you can do two simple things to help reduce rank leak:

If you have a page with lots of outgoing links, make sure that it also has links to the other pages in your site. You’ll be splitting the vote that way between outgoing and internal links, rather than passing all of it on through outgoing links.

Ideally, you want the page with the external links to be one with a low PageRank, reducing the outgoing votes. You can do that by minimizing the number of links from other pages on your site into the link page.

You could also use the rel=”nofollow” link attribute (see the section “Identifying nofollow links,” later in this chapter, for information) so that outgoing links don’t pass on PageRank.

There are other theories on how Google figures out which pages rank above others. One is the PigeonRank system described by Google in a technical document that you can find in Google’s technology area. Google suggests that “low-cost pigeon clusters (PCs) could be used to compute the relative value of Web pages faster than human editors or machine-based algorithms. By collecting flocks of pigeons in dense clusters, Google is able to process search queries at speeds superior to traditional search engines, which typically rely on birds of prey, brooding hens, or slow-moving waterfowl to do their relevance rankings.” Find details at www.google.com/technology/pigeonrank.html. No, really.

Page Relevance

The problem with PageRank is that it’s independent of keywords. The value is a number derived from links pointing at a page with no relation whatsoever to a specific keyword. A page may have a high PageRank, but that doesn’t mean it’s the type of page you’re looking for when you search for a particular keyword.

Thus, search engines add something else to the mix: relevance or context. The major search engines are attempting to do this sort of analysis by matching keywords. In effect, search engines are trying to create a context-sensitive PageRank or topic-sensitive PageRank. A topic-sensitive PageRank is dependent on a particular topic. Instead of counting any and all links, only links from relevant Web sites are included.

Page relevance is harder to measure, though. The concept of page relevance is that a link from a page related in some way to your page is more valuable than a link from an entirely unrelated page. A link to your rodent-racing site from a Web site that is also related to rodent racing is more valuable than, say, a link from your Aunt Edna’s personal Web site.

One way search engines are probably trying to do this is by using directory listings, such as Yahoo! and the Open Directory Project, to provide some context. Because Web sites listed in a directory have been categorized, it gives search engines a starting point to figure out what keywords and sites relate to what categories.

This discussion is getting complicated now, and you really don’t need to know the details. However, if you want to read a little geek stuff related to relevance or context, search for a paper, “Topic-Sensitive PageRank,” by Taher Haveliwala of Stanford University.

My feeling is that this sort of technology isn’t as advanced as many believe or as advanced as search engines want you to believe. Still, people in the search engine business swear that links from related sites are more valuable than links from unrelated sites. To be more precise, links from sites that a search engine thinks are related are more valuable than those from sites that a search engine thinks are unrelated. Because the technology is imprecise, search engines don’t always get it right. The fact is that no search engine really knows for sure if one site is related to another; it can only guess. As with everything else, relevance is just one ingredient in the mix.

Consider this example. Say you own a rodent-racing Web site. Among the links pointing to your site is one from a blog owned by an insurance agent. Now, most of the content on that site is related to the insurance business, but it just so happens that the owner is also a big fan of rodent racing. So is that link a relevant link? Of course it is! It’s from an insurance site but is placed there for good reason by someone who loves your site. That’s the real thing, the kind of link Google wants to see.

So don’t let anyone tell you that links from “unrelated” or “nonrelevant” sites have no value. A link from an unrelated site helps search engines find your pages, boosts PageRank, can provide information about what your site is about (if you have good keywords in the links, which I discuss later in the chapter), and may even bring visitors to your site. These links just won’t have the extra boost provided by relevance, but that doesn’t mean they have no value.

Hubs and Neighborhoods

Having said that relevance may be overrated, I still do recognize that relevance can be useful (it’s just not everything). The ideal link is one from a related Web site, not just any old Web site you can convince to link to you. The most powerful link hubs — networks of interlinked Web sites — are those that are tightly focused on a particular topic. Search engines respect that, want to uncover such situations and rank the target sites more highly, and are probably getting better every day at doing so.

Search engines are looking for what may be thought of as Web neighborhoods, communities, or hubs — groups of Web sites related to a particular subject, and the sites that appear to be most central to it. If you’re positioned in the middle of a cloud — a web of Web sites related to a particular subject — that’s a good thing.

Imagine a chart showing the world’s rodent-racing Web sites and their connections. In this chart, little boxes represent the Web sites and lines show links between the sites. (Figure 15-3 gives an example of such a chart.) Some of the boxes seem to be sitting by themselves — very few links are going out, and very few are coming in. Other boxes have lots of links going out but few boxes linking back. Now imagine your rodent-racing site. It’s a hub with many links going out, and many links coming in. That would look pretty important on a chart. Don’t you think search engines would find that interesting? In fact, search engines are trying to figure out this sort of thing all the time. Therefore, if you can build links to turn your site into a hub, that’s a great thing! (That doesn’t mean that all you need to do is build lots of pages with outgoing links, by the way — far more important are the incoming links.)

Trust in TrustRank

Yet another ranking concept: TrustRank. The idea, from a document written by Yahoo! search engine researchers, is that a small set of “seed pages” is selected manually; someone manually creates a list of Web sites and pages that are likely to be reputable pages that can be trusted.

Then computers measure how far each site is from the seed site, in the sense of how many links away it is. Sites linked to directly from the original list of trusted sites are assumed trustworthy, and sites they link to are probably reasonably trustworthy, and so on. The farther a site is from the trusted sites, the lower the site’s TrustRank.

The basic principle, then? Links from reputable sites are likely to be more valuable than links from sites that are not so clearly reputable. A link from CNN.com, for instance, is likely to count for more than a link from TheKentFamilyNewsService.com. (A link from CNN.com is likely to bring much more traffic from people actually clicking the link, too, which, people sometimes forget, is the original purpose of linking.)

Links from .edu (education) and .gov (government) Web sites are also likely to be more valuable because such sites are less easily manipulated for SEO purposes.

But remember, this is just one more possible way to weight pages. It certainly doesn’t mean that a site can’t rank well without lots of trusted links.

Inserting Keywords into Links

As Chapter 6 makes abundantly clear, keywords in your pages are very important. But keywords outside your pages can also affect the page results. That is, the keywords in links pointing to your pages are very important.

If you have hundreds of links around the world pointing to your site, with the words rodent racing in the links, then search engines will get the idea that your pages are somehow related to rodent racing. It actually makes a lot of sense if you think about it. If hundreds of site owners create links pointing to your site, and, in effect, say, “This site I’m pointing to is about rodent racing,” then Google is being given a darn good clue regarding what your site is about. In effect, Google has recruited site owners to tell it what other owners’ sites are about.

Link text, in geek terminology, is known as anchor text. The link tag is an <A> tag, and A stands for anchor. Thus, if you hang around in geek company, you may hear links referred to as anchors. For more on anchors, see the “From the horses’ mouths” sidebar.

In other words, Google and other search engines use links to get an idea of what the pages are about. Now, wait a second. This is important. If a link pointing to your site can describe to a search engine what your site is about, you’d better do all you can to make sure the link says what you want it to say! Even if you don’t own the site that points to the one you’re trying to optimize (or are trying to get another site owner or site manager to link to your site), it’s in your interest to get the right keywords into the link.

While you browse the Web, take a close look at links on the sites you visit. Now that you know how critical it is to add keywords to links, you’ll see that many links provide relatively little value to the sites they’re pointing to. Sure, a bad link is better than no link, but a bad link could be better than it is. Here are some of the problems you’ll see:

Image links, including buttons or banners linking to sites: Search engines can’t read images, so they’re not getting keywords from them. (You should add keywords to the ALT attributes in the image tag, but search engines may not value ALT text as highly as link text.)

Links containing words that are not useful keywords, such as company names: In most cases, company names don’t help you in search engines. Use the keywords your potential visitors and clients are using.

Combinations of descriptions and click here links: For instance: For more information on rodent racing — rats, mice, gerbils, and any other kind of rodent racing — click here. Wow, what a waste! All the keywords are there; they just have no link! Remember, click here links are a total waste of hyperlink space.

The Googlebomb lives

Just how powerful is putting keywords in links? Well, consider the Googlebomb.

The most famous example is the miserable failure Googlebomb from back in the years 2004 to 2007. Back then, if you searched at the major search engines for that term, the #1 result in all three was President George Bush’s bio page on the White House Web site.

This feat was accomplished by a small group of people using links in blog pages. Despite the fact that this page contained neither miserable nor failure, and certainly not miserable failure, a few dozen links with the words miserable failure in the link text were enough to trick the major search engines.

Eventually (after a couple of years), Google bowed to criticism and changed something to ensure that the president’s page no longer appeared in the search results. Yahoo! and Live Search (now Bing) did not make this change, at least at that time, so this particular Googlebomb, paradoxically, only worked on Yahoo! and Live. Today George Bush’s page still appears near the top of the first page of results on Bing and its partner, Yahoo!, but it doesn’t for Google.

I’m not sure what Google did to remove the George Bush result. It may have manually removed the site from the results, though it says that it “came up with an algorithm that minimizes the impact of many Googlebombs.” Nevertheless, Google didn’t completely throw out the concept of using keywords in links to tell it what the site is about. In fact, its statement suggests just that. It “minimized” the impact of “many Googlebombs”; it didn’t completely stop them.

How can you be sure? Well, I can prove it to you. Search Google for the phrase click here; what comes up as number 1? The Adobe Reader download page (http://get.adobe.com/reader). Why? Because millions of pages around the world have something like this:

The following PDF document can be read with Adobe Reader. To download it, click here.

Also, visit the Googlebomb page at Wikipedia (http://en.wikipedia.org/wiki/Googlebomb), where you’ll find information about all sorts of Googlebombs.

Note, by the way, that Googlebombing works in both Google and Bing — the Adobe site comes up high for click here in both search engines, for instance. However, there are differences between the search engines. The click here Googlebomb works on both well, but the miserable failure bomb works well on Bing but not on Yahoo!

Clearly, the way Googlebombs work has changed over time. Perhaps, for instance, in response to complaints about the original miserable failure search, Google changed its algorithm to say something like, “If we see x number of links of type y pointing to the site using keywords that don’t appear in the site, then ignore the links.”

In any case, though, the fact is the basic principle behind Googlebombing remains valid: Putting keywords in links tells search engines what a referenced site is about, and the more links with keywords the better.

I often have clients ask me why a competitor ranks so well, also stating, “Their pages aren’t better optimized than mine. We have more pages, and more pages with the right keywords. . . .” When I do a link analysis (see Chapter 16), I often discover that the poorly optimized, yet highly ranked, site has a huge number of incoming links, with just the right keywords.

Here’s the ideal combination for links and the pages they point to: The keywords in the link match the keywords for which the referenced page is optimized. If you have a page optimized for the phrase Rodent Racing, point links with the words Rodent Racing to that page.

PageRank versus Keywords

So which is more important, PageRank or keywords? Well, they’re both important, and the ideal link is a high-PageRank link with good keywords. But I want to make an important point. Many in the SEO business will tell you that low-PageRank links are worthless, but that simply isn’t true. Keywords in links are very valuable because they directly tell the search engines what the referenced page is about. So even low-PageRank links, if they have good keywords in them, can be valuable to you.

Good Links and Bad

Search engines must regard some types of links as more valuable than others. And it seems very likely to me that search engines will, over time, change how they regard links. They have, in fact, over time devalued certain types of links. For instance:

Links inside paragraphs of text are likely regarded as more valuable than links in large lists or links set apart from text.

Search engines could compare incoming links with outgoing links, and devalue a site’s links if it appears to do a lot of reciprocal linking from link pages.

Links pointing to a page inside a site might be valued more highly than links to the home page.

Outgoing links concentrated on a few pages might be valued less than links spread around a site.

When site owners figured out that links were valuable, they started playing tricks to boost incoming links to their sites. Some tricks were so egregious that search engines decided they were unacceptable. The trick you hear about most often is the link farm, an automated system that allows site owners to very quickly create thousands of incoming links by joining with thousands of other site owners to exchange links. Search engines don’t like link farms and will exclude link-farm pages if they identify them. Another trick is to create multiple shadow domains or satellite sites — groups of small, well-optimized Web sites that redirect traffic into a single site — and link them all into one site.

However, as with much in the search engine optimization business, another myth has arisen. You may hear that if a link is found to your site from a link farm, you’ll be penalized.

Let me set the record straight: Search engines do not penalize sites for incoming links. They can’t, or it would encourage dirty tricks. Want to push a competitor’s site down so that your site can beat it? Then link to it from as many link farms as you can. Obviously, it wouldn’t make sense for search engines to encourage this sort of thing, so links from such pages won’t hurt your site — though they won’t help it, either.

On the other hand, links to such sites really can hurt you. Because you do have control over links from your site to others, if a search engine decides that you are linking to a bad neighborhood, it may penalize you.

The bottom line is that you should avoid working with link farms because they can potentially harm you, and they won’t help you, anyway.

Do search engines ever penalize? Sure. However, with billions of Web pages in the large indexes, most penalties are automated (although Google also has a team that reviews sites for spam — often, sites that have been reported). To automate penalties, search engines have to create a very loose system that penalizes only the very worst offenses, or they risk penalizing innocent people. The proof is that if you spend time searching through the major search engines, you’ll find many pages that clearly break the rules yet are still included in the indexes. See Chapter 20 for more information.

Recognizing Links with No Value

Some links have no value:

If a page isn’t in the search-engine index, the links from the page have no value. Search engines don’t know about them, after all.

If a page is regarded as a link farm (described in the preceding section) or some other kind of bad neighborhood, as Google calls it, the search engine may decide to exclude the page from the index, or perhaps ignore the links coming from the page.

Some links to your site just aren’t links to your site; they appear to be, but they’re not. I explain in a moment.

nofollow links are a special form of link that usually don’t bring any value. I look at them later in this chapter.

Identifying links that aren’t links

How can a link not be a link? I’ve seen this happen many times. Someone gets a link to his site from another site with a high PageRank, perhaps a perfectly related site, and is excited. That’s a great vote for his site, isn’t it? Then some jerk, perhaps me, has to burst his bubble and point out that, actually, no, it won’t have the slightest effect on his site because it turns out that the link is not a link.

When is a link not a link? In cases such as these:

The link has been created in such a way that search engines can’t read it or will intentionally ignore it.

The link points somewhere else, perhaps to a program on someone else’s Web site, which then forwards the browser to your site.

Here’s an example:

http://ad.doubleclick.net/clk;6523085

This link passes through an ad server hosted by an advertising company called DoubleClick. When someone clicks this link, which may be placed on a banner ad, for instance, a message goes to a program on the ad.doubleclick.net ad server, which logs the click and then forwards the browser to www.yourdomain.com. You may think this is a link to www.yourdomain.com, and it may work like one, but as far as the search engine is concerned, it’s really a link to ad.doubleclick.net. This is also a common situation with affiliate programs through services such as Commission Junction; the links to your site go through the affiliate-service servers first, so they don’t count as links to your site. “But can’t the search engines figure out the redirect?” I hear you ask. Yes, but there’s more to this than meets the eye. One reason search engines probably won’t provide any value to redirecting links like this — and Google has stated that it generally doesn’t — is that such redirecting links are generally created for commercial purposes (advertising and affiliate programs) and thus are not true “endorsements” of the destination Web page.

Here’s another example: Suppose the person creating a link to your site doesn’t want the search engine to be able to read the link. This person may be trying to get as many incoming links as possible while avoiding outgoing links, so she does something like this:

<SCRIPT LANGUAGE=”JavaScript”>

<!--

document.write(“Visit <A HREF=’http://www.yourdomain.com/’>

Joe’s Rodent Racing site here</A>.”)

//-->

</SCRIPT>

The author is using JavaScript to write the link onto the page. You can see the link, other visitors can see it, and the link works when clicked, but search engines may not read the JavaScript, so they may not know there’s a link there. What appears to be a perfectly normal link on a Web page could be invisible to search engines; therefore, it may do your site no good as far as PageRank or any other link-popularity algorithm goes.

Here’s another form of link that search engines may not read:

<A HREF=”#” class=results onclick=”window.open(‘searchresult-

temp1.php?CS=cddzdzdrzfzpzpdc&SRCH=134893378&YD=0.88&RK=

3&PID=16&URL=yourdomain.com’,’merch’,’Height=’ +

screen.availHeight + ‘,Width=’ + screen.availWidth +

‘,left=0,top=0,scrollbars=yes,status=yes,toolbar=yes,

directories=yes,menubar=yes,location=yes,resizable=yes’,

false)”;>Everything About Rodent Racing!</A>

This is a real <A> link tag. However, it doesn’t use the HREF= attribute to point to a Web page. Rather, it uses a JavaScript onClick event handler to make it work. The JavaScript runs a program that, in this case, loads the page into another window. Again, the link works in most browsers, but search engines may not see it.

Incidentally, it’s also possible to do the reverse: make links appear to search engines to be links to your site when actually they’re links to another site. For instance, look at the following link:

<A HREF=”http://www.yourdomain.com/” onClick=”return

rewrite(this);” class=””>Joe’s Rodent Racing</A>

This link starts off well, showing yourdomain.com as the page being linked to. In fact, if you took the URL from the HREF= and pasted it into a browser, it would work properly. However, when someone clicks the link, the JavaScript onClick event handler runs, taking the domain and passing it to a JavaScript function called rewrite. Because search engines may not read JavaScript, they may not see what really happens when someone clicks the link. (In this example, the click runs through a program that tracks how many people use the link.) Search engines think it’s a link to your site, and I guess it is in a sense, but it has to pass through a program on a different site first.

Identifying nofollow links

There’s a simple way to tell search engines to ignore links; use the rel=”nofollow” attribute, like this:

<a href=”http://www.domainname.com/page.html” rel=”nofollow”>Link Text</a>

You can also block all links on a page by using the robots meta tag, like this:

<meta name=”robots” content=”nofollow”>

Tell the search engines not to follow the link — nofollow — and what exactly do they do? They ignore the link. It’s as though the link were nonexistent. They don’t bother following the link to the referenced page, and they don’t use the information to index the page. That is, you get no PageRank, Web Rank, or equivalent benefit, and no benefit from the keywords in the link. (That’s the theory, anyway; you can read my alternative theory in Chapter 18.)

More and more sites are using nofollow attributes to stop people from placing links purely to help their search engine rank. Here’s a classic example: Craigslist. The hugely popular Craigslist.com classified-ad site used to be a great place to put links; businesses would drop ads into the system and include links back to their sites to help boost their search position. So, in 2007, Craigslist decided to put a nofollow attribute into all links placed on the site in order to discourage this behavior.

Various browser plug-ins can show you nofollow links. For instance, after installing the Firefox NoDoFollow plug-in you can right-click a Web page and select NoDoFollow, and all the links on the Web pages you view will be colored; follow links will be gray, and nofollow links will be pink. (For this reason, nofollow links are occasionally called pink links.)

Recalling a Few Basic Rules about Links

I explain why links are valuable and how they can help boost your position in the search engines. But I cover a lot, so here’s a summary:

Links from sites related to yours are often more valuable than links from unrelated sites. (Because the relevancy software is almost certainly imprecise, this isn’t always the case.)

Links from pages with a high PageRank are more valuable than those from pages with a low PageRank.

Virtually all incoming links to your site have some kind of value, even if the pages have a very low PageRank or are not contextual. It may not be much of a value, but it’s there. (Assuming, of course, that the page with the link is indexed, that the link is not a nofollow link, that it’s not a redirect, and so on.)

The higher the number of outgoing links on a page that has a link pointing to your site, the lower the PageRank transferred to your site. This is because the vote is being shared. Thus, in some cases, a link from a low-PageRank page with no other links may actually be more valuable than a link from a high-PageRank page with many links.

Links are even more valuable when they include keywords because keywords tell search engines what the referenced site is about . . . even if the link is on a low-PageRank page.

A link strategy that positions your site as an authority, or a hub, in a Web community or neighborhood can be a powerful way to get the attention of the search engines.

And don’t forget that, as discussed in Chapter 7, even internal links — links from page to page within your site — are valuable because the keywords in the links tell the search engine what the referenced page is about.