The importance of metadata

Metadata is commonly defined as “data about data.” While this is a clever turn of phrase, it fails to encapsulate the purpose of metadata, which is to accurately describe content through summarization and unique identification. Just about every kind of digital document has metadata. Common image files like JPEG and PNG store color profiles, dimensions, and a thumbnail version. A Microsoft Word document contains page length, author, date of origination and last modification, word count, and more. A web page is simply another type of document, and it also uses metadata to describe its contents.

Images and word-processing documents have the benefit of third-party software adding metadata behind the scenes. You never have to count pages or manually add the date of modification to a Word document; Microsoft conveniently appends this automatically. Web pages, however, are built from scratch. The meta information has to be manually crafted, and its depth and breadth are dependent exclusively on the amount of effort the author wants to invest.

Documents with strong metadata have many advantages. First and most applicable, search engines like Google and MSN rely on this information to help index and order websites. However, it is used by a website’s internal search functions as well. As a pillar of information architecture, it is widely advocated by web professionals as a key component to proper development because it enhances the intrinsic nature of the document; whether the actual metadata contributes to SEO is less important than the web page being a good corporate citizen in the web community.

For web pages, the <meta> tag and all of its permutations are used for describing the metadata, the only exception being the <title> tag. These are all covered in the following subsections.

The page title. The title of the web page is arguably the single most important line of markup in the entire document. It is the text that appears inside the top chrome of the browser (in Figure 13-4 it is Search Engines - Google Analytics), which users refer to constantly, and is the default entry title when a site is bookmarked. In regard to search engines, it is usually the emphasized line in each entry of a SERP, although it’s not guaranteed, as you can see by the Yahoo example in Figure 13-5.

The title of the page should succinctly summarize the content. It should be relevant to the contents of the page, contextual to the rest of the site, and unique from any other page title on the site. Treat it as a unique identification mechanism. Adding keywords is important, but so is maintaining the human-readable factor; pages that use artificial or sensational titles come off as poorly as a sleazy used car salesman. Consider the homepage for our example company Rockstar Healthcare Staffing. A bad example would constitute non-semantic structure and word-stuffing:

<title>Traveling Nurse - Nurse Staffing - Healthcare ➢ Staffing -Rockstar Healthcare Staffing</title>

The overindulgence of keywords helps no one, least of all people looking for an honest company. Important keywords can be added just as effectively using normal language.

<title>Staffing Agency for Traveling Nurses: Rockstar ➢ Healthcare Staffing</title>

Getting the right keywords in the title is critical. Search engines rely on these to properly index websites, and people rely on them to identify the nature and business of the company. Earlier in the chapter we covered the three levels of keyword identification. Titles

should incorporate level 3 words and phrases without question, and level 2 if room permits. In the preceding title, there are three level 3 keywords: healthcare, staffing, and nurses, and one very important level 3 phrase: traveling nurses.

The order of keywords in a title can make a substantial difference in ranking as well. The most important words should be as far in front as possible. There is anecdotal evidence that ranking systems give more weight to the first few words, and people scanning a results page will latch onto words bolded to the left instead of the middle.14 There is an ongoing debate whether to include the company name—the general recommendation is that it should be present, but at the end, and that the stronger keywords should get bigger play at the front of the title tag.

The concentration of keywords is important because there is a finite amount of characters available to a page title. The W3C recommends a limit of 64 characters.15 Google displays about that without defaulting to an ellipsis. MSN and Yahoo are a bit more generous with around 70 characters, but this number seems to fluctuate (at one point, Yahoo was showing 120 characters), so make sure to stay within the safe zone of 60 to 64 characters, including spaces.

Description meta tag. Almost every document type that supports metadata enables the author to include a description. For web pages, this is a vital bit of content that can significantly impact rankings. However, unlike the <title> tag, the description is handled differently across the search engines—Google sometimes displays it in the results, others index the content but never display it, and still others ignore it completely. That being said, it’s still a vital slice of text that adds tremendous semantic value.

Like the title, there is a character limit, although it’s a lot more generous. Opinions of ideal length vary wildly, but the most practical measuring stick is Google’s results page, which displays 160 or so characters. Yahoo and MSN are not much different.

Descriptions should be written in normal, human-readable syntax with complete sentences and proper grammar. Keywords are important, but not the focus. The purpose of the tag is to describe the contents of the page in the simplest terms possible, not to fake out search engines with loaded clauses and fragmented phrases. For instance, this might appear in the homepage of Rockstar Healthcare Staffing:

<meta name="description" content="Rockstar Healthcare ➢ Staffing provides traveling nurses with high-paying ➢ radiology, orthopaedic surgery, and billing management ➢ jobs across the United States." />

The text is brief, descriptive, and not weighed down by ambivalence or unnecessary words. This focus on brevity and the inclusion of keyword phrases results in a potent description that not only talks about the page’s contents, but will be blessed by the search gods.

1. This evidence is anecdotal in the sense that just about everyone recommends it, but no one has any qualitative research backing it up.
2. www.w3.org/Provider/Style/TITLE.html

Keywords meta tag. Back in the day, the keywords meta tag was all the rage. People stuffed 1,000 characters worth of every iteration, theme, misspelling, and vaguely relevant term into a single tag, hoping search engines like Infoseek and AltaVista would reward the effort put into the giant tossed salad of language. Predictably, this practice quickly devolved into SEO “experts” adding Britney Spears, p0rn, Clinton, and other completely extraneous terms into the pages of honest (but naive) companies.

Needless to say, the keywords meta tag was beaten to the verge of death, and today exists only by the grace of developers unwilling to pull the plug on an old stalwart. For years, search engines have publicly denounced the keywords tag as a futile means of influencing ranking, and these days no search engine offers a might/possibly/snowball-chance-in-hell indication of supporting it. SEO experts will be the first to admit its demise—a panel of 37 industry leaders cited the keywords meta tag as one of the least contributive factors to positive ranking.16

However, adding the tag does not hurt, and on the off chance that it is supported in some infinitesimal state, or will be in the future, it will only benefit the content of the site to have that extra context. This does not mean overdo it—just add a few keywords that reinforce the title, description, and body text, like in the following example:

<meta name="keywords" content="rockstar, healthcare ➢ staffing, traveling nurses, jobs" />

This can be done quickly and without much thought. Many who are unfamiliar with contemporary SEO best practices place too much emphasis on the keywords meta tag, so make sure it’s used like garnish on a steak dinner—a bit of color not really intended for consumption.

Author meta tag. A common attribution to web pages is the author meta tag. This is not complex. Its only purpose is to define the author of the web page or site in general, whether an individual, the company as a whole, or a third party. The syntax is similar to the others.

<meta name="author" content="Rockstar Healthcare Staffing" />

Copyright meta tag. This is another optional tag, which defines the copyright information for a web page. This is usually the same across the entire domain; customizing it for every page is not necessary.

<meta name="copyright" content="Copyright 2007 Rockstar ➢ Healthcare Staffing" />

Robots meta tag. The robots meta tag is still in active use. It instructs how search engines are to index the content of the site, or prevents them from doing so. All spiders index content and follow links by default, so unless there is some reason to deviate from this standard, including this tag is redundant. We’ll cover it in more detail at the end of the chapter.

16. SEOmoz, “Search engine ranking factors, v2” ( www.seomoz.org/article/ search-ranking-factors).

URL structure

It should not be surprising that search engines place a lot of weight on the actual URL of a page. When a document’s physical location on the Internet incorporates words that match a search query, that page’s ranking is going to benefit from just the appearance of those terms. As you can see in Figure 13-6, Google bolds all matching terms—even ones in the URL.

The basic logic is simple. If a particular web page has a strong, descriptive file name, and it sits within a directory path of similar semantic vigor, and they are all hosted on a domain whose name indicates relevance to the query, search engines naturally think (usually quite correctly) that this location will be of interest to the user.

Page URLs that consist of database queries mean nothing to a search engine spider; they, like you and I, see only a sequence of nonsensical flimflam that looks like the Roman alphabet on acid. These strings are especially common in forums, FAQs, and other content-heavy places that are dependent on constant discussions between the server-side language and the database.

SEO benefits come when the web page’s path is intelligent, readable, and semantic. The following list represents different URLs that are all possible for a given page:

http://www.rockstarmedicalstaffing.com/web/srvcs/dindseg/ ➢

front/index.jsp?re=gihome67fssb

http://www.rockstarmedicalstaffing.com/srvcs/rad-assistants.html

http://www.rockstarmedicalstaffing.com/staffing-services/ ➢

radiology-assistants.html

The first is almost complete nonsense. The second is common in small sites, where unknowing developers cut corners with the file names. The last represents a site where the time was taken to design an unabbreviated directory structure and then fill it with files boasting equally sharp names. This not only makes the URLs easier to read when navigating the site, but increases the chance of matching some keywords in a future search engine query.

Beyond the actual words and directory structure, a few other small devices have been theorized to help in ranking. Most of these are unproven. People who engage in SEO as a profession run test after test to collect anecdotal experiences, but the search algorithms change so often that nailing any single advantage point is nearly impossible.

1. The fewer directories, the better: Many hypothesize that the closer a web page sits to the root of the site, the more weight it will be given by engines.
2. Hyphens over underscores: For multiple words, hyphens not only read better, but seem to score better results. Never use regular spaces—they are encoded as %20 and clutter the URL.
3. File extensions are irrelevant: The actual file type of the document (.html, .jsp, .asp, .php, .shtml, etc.) does not affect the ranking (nor should it).

Invisible page content

Sprinkled among the markup and the visible text are the small phrases and characters that fill in all the blanks between code and content. Class and ID values, alternative text for images and multimedia files, and link titles are all valid chunks of data, and can contribute to an SEO effort if used correctly.

Invisible content should champion the primary content. It needs to be relevant to the tag it’s supporting, but also work in context with the rest of the page. For example, if a web page is describing the benefits of a particular product, and the text is accompanied by an image, that image should contain an alt attribute whose primary function is explaining the image, not to be weighed down by a bunch of unrelated keywords. Take the following example:

<img src="/images/sweet-ftp-screenshot.jpg" ➢ alt="screenshot showing power of Sweet FTP on Mac OSX" />

The alternative text is short, descriptive, and uses a sprinkling of keywords to remain relevant to the overall message.

Many SEO experts will try to increase a page’s relevance by bumping the overall keyword density inside the HTML. In order to avoid altering the visible text, they load alt and title attributes with a mountain of keywords. In addition to being unproven, this tactic has drawbacks in terms of accessibility.

Visible page content

After the <title> tag, the actual HTML text that appears in the browser is the most important thing for developers, writers, and marketers to focus on. Optimizing the content may be the hardest part of SEO. A careful balance must be maintained between everyday text that people can read and understand, and text that has been obviously written to entice search engines.

The golden rule of SEO and the Web in general is simple: write for people, not machines. People write the text, create links, blacklist URLs, and contribute to websites; they are the drivers of the Web and search engines are simply riding along in the backseat.

That being said, there are several methods for constructing content to which search engines respond positively. None of these will make an asteroid-like impact on an SEO effort, but if used collectively and consistently, they can elevate a page far above a competing site that does not follow the same good advice.

Use headers. Text inside HTML is more than just paragraph and list tags. The language provides a series of header tags that are meant for visible page headings, from the primary document title to the lowest subhead. There are six header levels.

The first and most important is represented with the <h1> tag, and is used for the overall page heading. Theoretically, there should only be one on the page. Search engines place a lot of weight on the content, so treat it like the <title> tag—short, readable, keyword-friendly, and unique. (In fact, many recommend that these two tags should be nearly identical, since, in theory, they are doing the same thing, just in different places of the markup.) Do not abuse the <h1> tag. The name of the game is well-organized, semantic structure: documents should be properly ordered, like the chapters, headers, and subheads of this book. Having 20 top-priority headers on the same page just does not make sense.

After <h1>, the other headers go from <h2> through <h6>. Use them as appropriate. There is no golden ratio or perfect header pattern that triggers a better response from search engines; the earlier adage remains: write for people, and search engines will follow.

Push keywords to the top. When writing content, it’s important to weave keywords into the fabric of the text. However, when those words appear closer to the top of the page, search engines give them more emphasis and readers do not have to scan far to see if the content interests them. When both searchbot and human find key phrases in the first and second paragraphs, the assumption is that the rest of page’s content will relate to those themes as well.

This is more of an art than a science. It has to be felt out by the writer, and read correctly and naturally when published. A focus on level 1 and level 2 key phrases within the first few sentences will result in stronger optimization, especially if their primary keywords match the words used in the title of the page and the header tags.

Emphasize key phrases. Emphasizing key phrases and words within the text reinforces the page’s focus on these core ideas, and aids the reader’s eye in finding terms of interest. For instance, consider the following text. The bolded text emphasizes keywords the reader will be looking for.

Rockstar Healthcare Staffing has over 20 years of experience placing professional traveling nurses in a variety of healthcare positions across the country. We specialize in radiology, orthopaedic surgery, and billing management positions, with a focus on quality of talent and high-paying opportunities.

Always emphasize text with the <em> and <strong> tags. These have semantic value (“emphasis” and “strong emphasis,” respectively) that is understood by search engines. Never use <b>, <i>, or <u> tags—these describe only the visual display, and do not connote any additional meaning.

Provide descriptive outgoing links. A site that is regarded as a hub for the industry and ideas it represents will be blessed by search engines. Too many corporate sites are stingy with their outbound links, usually from the fear of people wandering off on some tangent and never coming back. The reality is that if people find your site interesting, they’ll find their way back.

Always give descriptive supporting text to outbound links. This is handled through the anchor text (the words between the <a> tags) and the value of the title attribute. This content boosts the ranking opportunities of the site benefiting from the link, which in turn will help the originating source. For example:

<a href="http://erl.wustl.edu/" title="Electronic Radiology Laboratory investigates digital imaging technologies for radiology departments"> Electronic Radiology Laboratory</a>

Building incoming links

Above all else, there is one primary tactic for building search engine karma: incoming links. Google was revolutionary in the respect that its PageRank system largely based its results on the popularity of the site, which essentially boiled down to how many links were pointing to the domain. Since then, every search engine has copied this model.

Today, the secret ingredient to Google’s ranking system is mired in millions of lines of highly secure algorithmic code, stored on black boxes in rooms where only the most privileged employees have access. There is not much the world knows about this secret sauce. But the one thing experts, pundits, and casual passersby do know is that a website’s ranking is based on more than just the quantity of incoming links—there is a subset of tests that determine each external link’s value:

1. The anchor text: This is the text that is contained between the <a> tags. It should be relevant and descriptive; for example, <a>staffing for nurses</a> is a lot more valuable than <a>click here</a>.
2. The title attribute: The title attribute describes the link in question; it reinforces the anchor text with a phrase that lets the user know exactly where they are heading—for instance, <a href=" http://www.rockstarmedicalstaffing.com" title="Rockstar Healthcare Staffing provides staffing for traveling nurses">staffing for nurses</a>.
3. Context of the link: A link sitting within relevant material is seen as more valuable than one floating out in space. Context also goes beyond surrounding text. For instance, a link coming from a medical staffing blog or directory is a lot more valuable than a link coming from the designer’s portfolio site: the first is pertinent, the second is peripheral.

1. The search engine value of the referring site: If the site does not perform well in search engines, it is not recognized as an influence, and its link is worth less than that of a high-ranking referrer. This is easiest to qualify in terms of PageRank. A site with a PageRank of 5 is worth far more than a PageRank of 4, and getting a few links from sites with PageRank of 6 or more can do wonders.
2. Age of link: Believe it or not, search engines look at not only the age of the site as an indication of authority, but also the age of the link.

Not every link is going to be perfect. In fact, the chance of meeting all of these criteria is slim, and there are times when you will have to settle for whatever you can get. A lot of companies may not have immediate access to highly ranked sites from which to aggregate incoming links, so they will have to start small, building a referral network through some footwork. The following subsections give some places to start.

Directories. Directories are a great place to start in an SEO effort. There are many to work with, they are generally reliable, and the message can be controlled by the submitting company. Directories are manually edited by a real human, meaning that duplicate or spammy sites do not get listed, but it also means that the submission process can take awhile. Some directories are free, while inclusion in others requires payment.

The most well-known directory is Yahoo, and was the company’s founding model until it became apparent that search was the future. There are a zillion other directories on the Web,17 each claiming some type of niche or specialty, and almost all are pining for submissions in an effort to build their own search engine status.

For SEO campaigns, there is one place to start: dmoz.org, shown in Figure 13-7. This is the home of the Open Directory Project (ODP), where an army of editors oversees the largest collection of website listings under one roof. Dmoz is commonly used by major search engines for descriptions and other information—in Figure 13-5, both MSN and Yahoo source www.nhai.com’s description from the ODP, not the description provided in the site’s metadata.

Press release and article sites. Almost every company produces some volume of content that is intended for public distribution. This includes articles, press releases, how-to instructions, and more. Traditionally, this content has been relegated to the corporate site, waiting for people to stumble upon it through a search. While it is good practice to host this material, distributing the text via the broader Web helps cast a net of content that not only builds interest and name recognition in the authoring company, but builds a network of incoming links as well.

17. www.seocompany.ca/directory/free-web-directories.html is a good list of directories; www.best-web-directories.com is also a good source.

Take press releases. Certainly every public company writes and publishes them—they are a staple of the investment media’s diet. While it’s a good practice to publish press releases on the website for casual browsers to discover, and while there is a chance a release may get picked up by an online or printed publication, the wait-and-see model is anything but efficient. Complement this effort with proactive publishing on external press release web-sites, such as the following:

PRZOOM ( www.przoom.com)

Free-Press-Release.com ( www.free-press-release.com)

1. PR Leap ( www.prleap.com)
2. Press Method ( www.pressmethod.com)
3. OpenPR ( www.openpr.com)
4. ClickPress ( www.clickpress.com)
5. UKPRwire (UK only) ( www.ukprwire.com)
6. Pressbox (UK only) ( www.pressbox.co.uk)

Some of these require the creation of an account, but almost all will allow a link back to the corporate homepage. These are positive incoming links. The content is relevant to your site, and most of these sites rank well in search engines by themselves.

In addition to press releases, a company might staff writers that regularly produce industry articles, from industry commentary to breaking news to instructional media. Although this is great content for the host site, there are many third-party websites dedicated to republishing article content; authors can submit their material to be picked up by other websites. Full copyright is retained by the original writer, and each article is accompanied by a link back to the corporate website every time it is republished.

For instance, say you worked for Rockstar Healthcare Staffing, and you wrote an article offering advice for nurses thinking about signing on with a staffing agency. This is great content, applicable industry-wide. You publish it on Rockstar’s website, and it attracts a few visitors from Google. Seeking to take advantage of the content, you submit the material to a few article sites like GoArticles18 and Article Dashboard,19 and the piece gets picked up several times, resulting in a pile of links back to your domain.

The math is simple. If an article is submitted to one site and gets republished 25 times, that’s 25 links back to your site:

(1 article) ✒ (1 site) ✒ (25 reprintings) = 25 links

It does not take much imagination to make that number grow exponentially:

(10 articles) ✒ (10 sites) ✒ (25 reprintings each article) = 2500 links

Considering there are literally hundreds of article websites,20 this can be a very effective way to quickly create a network of inbound links.

The industry circle. Few companies operate in isolation. Every industry has its collective of forums, directories, news boards, media sites, alliances, blogs, and more. Look around at the sites people look to for expertise—the centers of information—and see if they offer the opportunity to link to related companies.

The first place to look is within your own industry alliances. Many businesses list their strategic partners on their website in a centralized directory, as you can see in Figure 13-8. These listings are almost always free—provided for the benefit of the visiting prospect— but with the expectation that a reciprocal link will be provided.

When marketing within the industry, always think about ways to publish a link. For instance, when posting to a forum, add a signature with the corporation’s link. Many forums are ranked highly because of their cycle of fresh and unique content, and these small but numerous links can add up over time.

1. www.goarticles.com
2. www.articledashboard.com
3. www.wilsonweb.com/linking/wilson-article-marketing-1.htm has some of the top article sites, and www.styopkin.com/article_submission_sites.html has a list of over 500 more.

Blogs are one of the best mediums from which to receive incoming links, but like traditional media, there’s no easy way to solicit for these. In order to get the tenuous attention of bloggers, the company has to do something worth blogging about, either good or bad—new product releases, viral marketing efforts, and poor customer service can all elicit commentary. (Try to avoid negative discussion. You do not, after all, want to rank highly for the term “poor customer service.”)

Links from other sites within a company’s industry count for a lot. No one knows how much emphasis search engines place on site-wide topicality, but the suspicion—if nothing else—is that the significance will only grow in the future.

Submitting to search engines

While a lot of effort goes into building incoming links, not much can be done about actually submitting sites to search engines the old-fashioned way. MSN and Ask do not even provide the opportunity to manually suggest a URL; they rely on their searchbots to find and index websites. Because of this, the single best way to get sites listed in any search engine is to build the network of incoming links.

Google and Yahoo, on the other hand, still provide the opportunity to submit a website.21 There is no guarantee that a site will be picked up from this effort, but it certainly cannot hurt the process. The key is to follow the directions for each explicitly—the systems are sensitive to spamming and will mercilessly blacklist a site that even smells of devious technique. For Google especially, it is a very good idea to also create an XML site map, as explained in Chapter 4.