What Search Engines Look For

As you learned in Chapter 10, "Get Your Site Indexed," it can be a challenge just getting your page into the search index. When you do, you face your next challengewhat the search engine thinks of your page. An organic search engine categorizes your page according to dozens of criteria, some driven by explicit tagging, but others based on judgments the search engine makes from an analysis of your page. We investigate the most important criteria search engines look for, which they use to make two different kinds of decisions:

• Filtering. For each query, search engines decide which pages are in the search results list and which are not. If your page meets the filtering criteria, it will be in the list somewhere.

• Ranking. For each query, search engines sort the search results by relevance to decide which pages are shown at the top of the list. The better your page, and the closer it matches your query, the higher it will rank in the results.

What you put on your page is your best chance to influence the search engine's decisions as it filters and ranks the search results for each query. We check out filtering now.

Search Filters

Searchers use search filters to set their search's scope. Pages that are not included by the filters for a query do not appear in the results. For example, a searcher using the Yahoo! Australia and New Zealand site can choose to search the entire Web, or just Web pages from Australia or those from New Zealand. If the searcher's query for "digital cameras" limits results to pages from Australia, no pages from outside Australia will be shown, regardless of how closely they relate to the "digital cameras" search, because they are excluded by the Australian country filter.

The two most important search filters are for language and country, but we look at others, too.

Language Filters

Big news! People like search results in the languages they know! Okay, that might not be the lead story on the nightly news, but it is very important for search marketers. When a searcher enters a query as shown in Figure 12-1, only pages written in Japanese will be shown in the search results.

Figure 12-1. Language filtering in Google. Searchers can choose to limit results to their local language.

In some countries, such as Japan and China, the vast majority of searchers want their results limited to their native languages; in other places, however, such as Sweden, searches can be conducted in Swedish or English. Searchers in different countries have different preferences.

For the search marketer, what is important is that the search engines know the language of your page. If your page is not correctly identified, it will be missing from searches that should include it, lowering your referrals.

So how do search engines decide the language of your page? There are several different methods:

• Language metatag. Many Web pages contain an HTML tag declaring the language of the page (such as <meta http-equiv="content-language" content="ja"> for Japanese). It sounds simplethe spider reads the tag and the search engine knows the language, right? Not so fast. A very high percentage of the language metatags are flat-out wrong. The tags are missing, they are syntactically incorrect, or they have the wrong language encoded. Search engines do look at this tag, but they never decide the language of the page based on this tag alone.

• Character encoding. Computer files (including HTML pages) require a key to correctly interpret the characters (letters, numbers, and so on) in the file. That key is called the character encoding, and is declared on Web pages in a metatag (such as <meta http-equiv="content-type" content="text/html; charset=shift-jis">). Web browsers and spiders assume the page is encoded for Western languages, so pages written in those languages do not require this tag. Asiatic, Arabic, and Cyrillic text does require the tag for proper display by browsers; so when the search engines see the tag for these languages, it provides high confidence as to the correct language of the page.

• Content analysis. Search engines make their final determination of the page's language by studying the character patterns in the content. The correct language can be detected with very high accuracy for pages with as few as two dozen words, with metatags being used only for pages where the language is unclear after analysis.

For the most part, search engines will correctly detect the language of your pages without any action on your part. For pages with very few words, it is important that the language and character set metatags on your page be encoded correctly to ensure that your pages are identified in the proper language.

Country and Region Filters

Often, searchers do not want to limit results to a languagethey want all results within a particular country. This is particularly true of transactional searchers in the Buy stage of the Web Conversion Cycle. They want to buy from a vendor in their country that uses their currency and will not charge a king's ransom to ship the item.

Limiting by language does not do that. German pages exist in Germany, Austria, and Switzerland, French content in France and Canada, Spanish pages in Latin America as well as Spainyou get the ideaand English content all over the place. So, most search engines apply filters by country or by region. Searchers can always use the Advanced Search interface to specify these filters, but relatively few do. Instead, most local searches have a default filter, or allow selection between two or three filters on the search page, specifying a particular country, region, or language.

So how do search engines know which country your Web pages are from? They look at where the page is hosted and they examine the URL itself. Every Web page has a URL, and the domain (www.company.com) is resolved to an IP addressa unique number that points to a server somewhere on the Internet. Search engines can use that IP address to determine the country in which that server resides. Web pages hosted on servers within each country are part of that country's filter.

But pages hosted outside a country can be included in a filter, too, if they are named appropriately. Any URL that ends with the domain for a country is included in that country's filter. A British company might have a URL of www.britishcompany.co.uk (for example). The co in the name indicates it is a company, and the uk that it is a British company. So, a search within the UK brings back Web pages that end in ukwhat technical types refer to as the "top-level domain." Some countries just use their country code (such as de for Germany) with no co to indicate companies (such as www.deutschefirma.de). It sounds logical, and it worksfor Web sites that do business in a single country.

For the global search marketer in a multinational company, however, country (and region) filters can prove problematic. You want the country content on your site to be included by search engines when they filter by country, but it is not always easy to do, because of the way that the search engines' country filters work. Your multinational site is probably hosted centrally and uses com as its top-level domain, no matter what country the content is for.

Let's look at an example. Microsoft, like most multinational companies, maintains country "sites" that are part of its com top-level domain. Entering the URL www.microsoft.de redirects you to www.microsoft.com/germanywhich does not qualify as a page from Germany. You see, according to the search engine's country filter rules, all pages considered to be from Germany must be hosted locally or have a top-level domain of deMicrosoft's page is hosted centrally and has no content under its de name, so the search engine indexes the content from its com domain (the www.microsoft.com/germany page). Later, when searchers limit results to pages from Germany, this page will not be found (because it is a com and not a de and is not hosted in Germany); when they limit to German language pages, however, it will be found (because it is rightly analyzed to be written in German).

What is the global search marketer to do? First, don't panic. Many searchers understand this problem and regularly toggle between language and country filters to get what they want. If your target customers are not terribly sophisticated Web searchers, however, you might want to approach your Webmaster about changing the way your site is organized so that your country pages do use the top-level domains for each country. Or you can ask that your country pages be hosted at IP addresses within each country. Your Webmaster is unlikely to relish these suggestions, because they make your Web site harder to manage, but in the short term you might have no alternative.

In the long run, you should expect the search engines to address this issue. They are painfully aware of this problem and are taking some steps to ameliorate its impact. Some of the larger search engines already use your site's IP address to see whether your pages are hosted within the correct country, so that can help some of you. If you cannot adapt your site to use the proper top-level domains, we suggest ensuring that your URLs and content strongly reflect the country of the page. Microsoft's approach of adding germany (or de) to the URL while also placing the words Microsoft Deutscheland at the top of the page (itself written in German) might someday be enough clues for search engines to accurately discern the page's proper country. If you cannot satisfy what the search engines are currently looking for with their country filters, at least prepare your content to be as ready as possible for what they might be looking for someday. In addition, the better your content, the more likely it will draw links from sites that are included in the country filterfor some search engines, enough high-quality links from country pages can get your pages recognized as country pages, too.

Other Filters

Search engines offer searchers other filters that might sometimes be important to search marketers, but each search engine provides a different set.

Most search engines enable searchers to filter by type of content. Most have some kind of "picture" or "image" search, some can filter news stories, and most have an Advanced Search interface that filters by document type (such as Adobe PDF files or HTML files).

You might think it is valuable to be the #1 PDF file in Advanced Search, but it really is not, because so few searchers will take the extra time to use Advanced Search. In general, the more clicks required to execute a search, the fewer searchers will do so.

So before you get excited about these specialized searches, think like a searcher. If you work for a news organization, searchers for your site might make that extra click on the News tab in Google, so having high rankings for news stories could help you meet your search marketing goals. Image searches might be important to a seller of fine art prints. For the most part, however, none of these specialized searches are of much interest to the search marketer. (Some search engines offer a tab for shopping search, which might be important to youwe cover that in Chapter 14, "Optimize Your Paid Search Program.")

Most search engines enable searchers to set preferences that control how all of their searches work. Most preferences are unimportant to search marketers, such as the number of results on a page or whether search results open a new browser window, but one can be very important, because this preference is a filter.

The so-called Adult Content filter suppresses pornographic or otherwise sensitive material from the results. The issue for search marketers, as you might expect, is how accurate the filter is. In the past, news reports have trumpeted breast cancer sites (for example) suppressed by such filters, but modern search engines generally do a good job on these filters because of their strong text-analysis capabilities. Most search engines, by default, filter out just the most egregious scatological and sexual content, while leaving in explicit scientific and informational content. A strict setting can be chosen by searchers as their preference; if so, setting the preference just once then employs the strict filter forever. Search marketers whose site might contain sensitive material might want to monitor their page rankings so that they are not unfairly filtered, and might want to police word usage on their sites to avoid being filtered. Pay special attention to message boards on your site frequented by your visitorsif your visitors use inappropriate language in their posts, your site might be snagged by this filter.

You have finished your grand tour of search filters. Depending on the nature of your site, search filters might be critically important to your efforts, or you might not have to think about them much anymore. But now it's time to pay attention, as you learn what is behind a search engine's ranking algorithm.

Search Ranking Factors

A ranking algorithm is the mathematical formula a search engine uses to score pages against the query to see which pages are the closest matches. But what goes into that formula? How can your pages get consistently high scores for your targeted keywords?

It's time to answer those questions. Chapter 2, "How Search Engines Work," explored the basic concepts behind search ranking, but in this chapter we go deeper, explaining more of what search engines are looking for and, later in this chapter, showing you practical ways to help your pages score high. As you read this, keep in mind that the highest rankings mean nothing if your pages are excluded by the filters listed abovestrike out on a filter and your page is out of the results list no matter how closely it matches the query.

If your page is included by the filters for a particular query, the ranking algorithm takes over, looking at every page containing those words and deciding how your page stacks up against the others for that query. There are no right or wrong answers from the search enginethe engine tries finds the highest-quality pages matching the query. The ranking algorithm contains many factors, components that are scored for each page. If your page scores the most points, according to the ranking algorithm's factors, it will get the #1 slot in the results.

In Chapter 2, we discussed how complicated a search engine's ranking algorithm is. Because a ranking algorithm is such a closely guarded secret, no one can publicly state how many different factors a ranking algorithm weighs, but some say there are more than 100. Clearly, not all 100 factors are equally important, so we concentrate on the more important factors here.

Ranking factors come in two main varieties:

• Page factors. Web search ranking algorithms rely heavily on components that have nothing to do with the query entered, such as the strength of links zo the page, the number of visits to the page, and many more. These factors boost a page in the rankings for searches on any word that occurs on that page.

• Query factors. As you might expect, the particular query the searcher enters weighs heavily. The number of occurrences of the words in the query, where they are found on the page (title, body, and so on), and many other elements are weighed by the search engines when ranking results.

You can optimize your content for both page factors and query factorsneglecting either one will derail your search marketing program. We investigate page factors now.

Page Ranking Factors

The moniker "page factors" is a bit of a misnomerAndrei Broder, a Distinguished Engineer at IBM Research, prefers to call them query-independent factorsbecause they are not so much about the page as not about the query. So-called page factors can take into account anything the search engine knows about the page itself, the pages that link to that page, the site that contains the page, and many other components. What this means is that any particular page's page factor score is exactly the same for every querya page with strong page factors starts out with a high score for every word that is on that page.

Every Web search engine uses page factors as a critical component in its ranking algorithmGoogle's PageRank is the most famous example. As explained in Chapter 2, when a searcher enters a broad query, such as the word "camera," the search engine needs a way to decide which few pages, among the millions that contain the word camera, are the ones to rank at the top of the list. The pages with the most occurrences of camera are probably not what searchers want; they want the most definitive pages. Only page factors can make that determination. Let's look at the most important page factors in a search ranking algorithm:

• Link popularity. As discussed in Chapter 2, Web search engines put great stock in your page if other Web sites link to your page for the subject of the search. For example, the Snap Electronics digital camera product category page (the hub page that we first saw in Figure 7-2 in Chapter 7, "Measure Your Search Marketing Success") has many links to it that all have the anchor text (the clickable words in the link) digital cameras or SnapShot digital cameras. That causes the page to rank highly for the queries for "digital cameras" and "snapshot digital cameras." Although link popularity is a critical page ranking factor, it does not have much to do with how you optimize your content (because anchor text in the links is on someone else's site, not yours), so we hold off on the in-depth treatment of links until Chapter 13, "Attract Links to Your Site."

• Popularity data. Now that many searchers use search toolbarsthe Google and Yahoo! toolbars are the most popular, but many other search engines have them, toothey can gather information about which pages are visited the most. Unbeknownst to many toolbar users (although openly spelled out in their terms of use), the search engines keep track of which pages searchers are visiting, even when they are not searching. Ask Jeeves, for one, confirms using popularity data to give more popular pages a "boost" in the rankings when searchers look for words that appear on those pages. Other search engines might be doing the same.

• URL length and depth. You learned in Chapter 10 that dynamic URLs with more than two parameters might cause spiders to avoid the page, but you should know that longer URLs, in general, reduce your page's ranking in a small way. Search engines are more likely to "boost" pages that are closer to the root directory of your site, so the same page located deep in your site (perhaps at www.yourdomain.com/news/announce/today/top.html) might rank a bit lower than if placed nearer to your home page (say, at www.yourdomain/news/top.html).

• Freshness. If your page has not changed in a long time, its rank might be reduced because search engines suspect its information is out-of-date.

• Page style. Pages that are grammatically correct allow the search engine to better score relevance. Pages that are organized like a newspaper article (important words at the top, somewhat repeated throughout, and reinforced at the end) are sometimes said to have an advantage. In many cases, folks who use unethical spam techniques are flagged by these factors because their content is written in a stilted way to repeat keywords ad nauseum.

• Site organization. One of the best ways to strengthen your page factors is good Web design. Work with information architects to make the site simple to navigate with a well thought-out linking structure. Use meaningful words in your URLs, but do not take the spam route by stuffing in three or four keywords between hyphens. Use as simple a page layout and design as possible.

• Spam-free. Every page on your site contains text that can inform a visitor and attract a search engine. When you try to mislead a search engine with spam techniques, your page will be penalized in ranking, possibly leading to your site being banned.

As mentioned previously, although it is easy for us to think about these factors as relevant to just one page, search engines are more sophisticated than that. They look at links to your whole site, not just one page. They check for profanity on your whole site, even if most pages are "clean." If most of your pages are updated frequently, do not obsess about changing your "History of the Company" page every two months.

STUPID CONTENT TRICKS: HOW TO GET YOUR SITE BANNED IN FIVE EASY STEPS

Tricking the search engines to rank your pages higher than they ought to is called spamdexing, or simply spam. Throughout this book, we have warned you about a host of spam techniques, because spamming can get your site banned from search enginesa nasty wrench thrown into your search marketing plans.

There are many different ways of tricking search engines, people being the clever creatures they are, but we cover only the content spamming techniques here. Our goal is not to teach you how to perform these techniques. Rather, you should know enough to spot them to prevent your site from running afoul of the rules and suffering the consequences. Or enough to discover them on a competitor's site so that you can justifiably turn them in.

• Doorway pages. Any page that is designed solely to achieve high search rankings, but otherwise has no value to visitors to your site, is a doorway page. Search landing pages are not doorway pages, and we talk about the difference later in this chapter.

• Keyword stuffing. Also known as keyword loading, this technique is really just an overuse of sound content-optimization practices. It's good to use your target keywords on your search landing pages, and use them often, but when you start throwing them in just to attract the search engines your pages can be flagged. Dumping out-of-context keywords into the alternate text for images, or into <noscript> or <noframes> tags, is a variation of this same unethical technique.

• Hidden text. HTML offers many opportunities to place text in front of the spider that the visitor will never see. Displaying text in incredibly small sizes, or with the same font color as the background color are hoary spam techniques. Newer approaches include using style sheets to write keywords on the page that are then overlaid by graphics or other page elements. In short, any time you can see text in the HTML source of a page that does not show up when you view the page in your browser, it is probably spamthe only exception is valid HTML comments, which the browsers and spiders both ignore.

• Duplicate tags. Using duplicate title tags or other metatags have been rumored to boost rankings in times past. The same style sheet approach that can hide text can also overlay text on top of itself, so it is shown once on the screen but listed multiple times in the HTML file.

• Duplicate sites. Why stop at duplicate tags when you can clone your whole site? You duplicate the content in slightly different form under several different domain names and then have each of your sites link to each other (to increase their page ranking factors). Maybe your sites can grab six slots in the top ten results.

And the really bad news about all of these techniques is that sometimes they do work. Search engines do get fooledusually by people more industrious and harder working than us. Most of the time, however, spam techniques are like stock tips. When you hear the tip, it is probably too late. The stock price has already gone up and the search engines are already implementing countermeasures.

As you read the list, you might notice that you do not have a lot of control over some of these factors, and it is true, in general, that page factors are harder to influence than query factors. But you are not helpless. Although you cannot directly affect your site's popularity, for example, you can indirectly affect it in many waysthrough search engine marketing, by attracting more links (as explained in Chapter 13), and many other ways of getting attention.

Query Ranking Factors

For you control freaks out there, start salivating. The query-dependent ranking factors, which we call query factors for short, are what you will spend most of your time on, as you lovingly craft each search landing page to best appeal to search engines, and (do not forget) the searchers themselves.

Page factors are constant across every query. A page with high-scoring page factors takes that score with it for every query. And although page factors are important, there must be something going on that is query related, or else the same pages would be at the top of the search results for every query. There is something going on the query factors.

But before talking about query factors for ranking, there is one filter that we did not address back in the filtering section, because it makes more sense to discuss it now. A very powerful filter is used on every queryat least one (and typically all) of the words in the query are expected to be found on your page. If none of the words are on your page, that page is filtered out of the results list, no matter how wonderful its page factors are.

Now that sounds simple and obvious, doesn't it? Except that it is not precisely true. One exception applies to that filtering rule: if enough pages link to your pageif they link using the query words in the anchor text of their links.

Let's look at an example. Figure 12-2 shows the search results for the word "laptop" in Google. Apple has the #3 result, but the word laptop does not appear anywhere on the page. It must be that there are so many other pages linking to this one (containing the word laptop in the link) that Google is convinced (correctly) that this is a good result for that query, even without any occurrences of the word on the page itself. If Apple placed laptop on this page, maybe it would rank #1.

Figure 12-2. The power of page factors. Once in a while, the links to a page drown out the ranking factors from the content on the page itself.

That does not happen very often, however. The vast majority of the time, your page must contain the keyword to rank highly for that query. When your page gets past that filter (and the other filters we discussed earlier), it is in the results list and ranking takes over. Each page in that list comes with its predetermined page score, such as Google's PageRank, the score associated with that page based on its cumulative page factors. Pages with high page factors get a head start in the scoring. But then the query factors take over to decide the winner.

Some query factors can apply to any query, whereas others kick in for multiple-word queries only. Here are the universal query factors:

• Keyword prominence. All words do not have equal importance on a page. Words in a title or in a heading are more important than words in a body paragraphthese locations of keywords are their placement. Keywords also show their value by their positionhow close to the beginning of a page element that they are. For example, words at the beginning of the body (or the start of the body) are usually more important than those that show up later in that same element. So, when we combine the concepts of placement and position, the most prominent keyword location is the first word of the page's title. Search engines look for a keyword in prominent places (and in prominent positions within those places) because it is one of the best clues to what the page is actually about. Pages with a keyword in prominent locations tend to be good matches.

• Keyword density. What percentage of the totals words on the page are the occurrences of the keyword? That is keyword density (sometimes called keyword weight). In Chapter 2, we mentioned that search engines look for around 7 percent keyword density, but that is just a guess. Your page will not be sent to the spammer's graveyard if you have 14 percent density, but it might not help you rank higher either. Density helps ensure that a page with 7 occurrences in 100 words is not outranked by a PDF with 12 occurrences in 5,000 words.

• Keyword frequency. As discussed in Chapter 2, search engines at one time placed much more stock in keyword frequencythe number of times the keyword actually occurred on the page. Given spammers' predilection for keyword stuffing, ranking algorithms began to favor keyword density. But frequency still matters. You cannot stick a page out there with the single word camera in the title and the body and expect high rankings for the query "camera." Yes, it has wonderful prominence and 100 percent density, but only two occurrences of the keyword. To get a top ranking for a keyword with strong demand, frequency is still important.

• Query intent. As search engines apply more and more analytical firepower to your pages, they are working to match the searcher's intent (such as navigational, informational, and transactional) to pages that satisfy that intent. Although no two search engines treat searcher intent in identical fashion, search engines generally respond to navigational queries with site home pages, whereas informational queries yield pages with several hundred words on them. Certain queries contain clues about the type of documents to returna search for "maytag jetclean dishwasher manual" might return the PDF product manual or a page with a list of links to the PDFs, for example. Search engines are always looking at new ways to improve recognition and satisfaction of searcher intent, so the more accurately your content reflects its purpose, the better your edge.

• Contextual relevancy. The newest frontier for search engine relevancy is the searcher's context. Context includes components that are permanent or semi-permanent, such as gender, job role, and marital status. Context can also include more ephemeral factors, such as current geographic location, the subjects of pages viewed recently, and recent search keywords. Currently the major search engines try to guess a searcher's physical location so that queries for "hardware store" will see ones near them. Other contextual factors might become important in the future. Those search toolbars that collect page popularity information can associate recent page views and searches with the searcher, although none have claimed to put it to use for ranking as of yet. Similarly, Yahoo! offers many services that require registration, so it can collect more permanent information about the searcher. All of this information could be put to use at some point to improve relevance, if searchers do not object due to privacy concerns. Such "personalized" searches have the potential to complicate rank checking for search marketers by changing the whole idea of "ranking #1" to "ranking #1 for a certain percentage of people," but search engines are taking mere baby steps in this direction so far.

You can see that every query undergoes complex analysis, and your pages do as well. But queries containing more than one word are evaluated in an even more complicated way, because of the interplay between the words. Take a look at the factors that apply only to multiple-word queries:

• Term rarity. When queries contain more than one term (words are called "terms" in search parlance), the search engine wants to know which words are the most important across the Web, because that helps the search engine find the best pages. Consider the query "hotels in london" from the search engine's point of view. A search engine knows the frequency of every term's occurrence across the Web. The term in occurs on most English pages and is therefore not a "good discriminator"it does not help the search engine pick the best pages for the query because pages containing the word in are just about as likely to be good answers as pages that do not. Rarer is the term hotels, but it is not as rare as London. For each term, the engine calculates its inverse document frequencya logarithmic formula that produces very high values for rare words and very low values for common words. The search engine then takes each of the three terms and executes a formula for each page in the search results known as TF*IDFnormalized term frequency (what we have called keyword density) multiplied by inverse document frequency. This compli cated math tells the search engines which pages have the highest keyword densities for the rarest terms in the query.

• Term proximity. As discussed in Chapter 2, the best pages contain all the terms in the query right next to each other in the same order they were listed in the query. So, pages containing hotels in London might be the best. But the search engines apply more judgment than that. Because the term in is so common, for this query pages containing "london hotels" might be just as good as those with "hotels in london." Other queries that contain no common words might emphasize word order more. In all cases, having all the words close to each other is a good thingcertainly better than a page that has numerous occurrences of London and hotel separated by several words or sentences.

Remember that search experts spend their whole careers crafting and polishing these formulas, so no short explanation will give you a complete understandingand you do not need one. You do need to understand the basics of what search engines are looking for, however, and now you do.

Although we have dealt with page factors and query factors separately, to make them simpler to explain, on every query the search engine mixes them together to derive the best ranking for the results listand different queries emphasize one set over the other. For example, imagine a query such as "digital camera." There are millions of occurrences of those words, and keyword density does not help much in determining the best pages. So page factors become critically important in deciding the top ten. But for the query "maytag jetclean dishwasher manual," relatively few pages contain all of those words, especially in proximity to each other, so query factors probably drive the top results more than page factors. Obviously any page that excels in both, for a particular query, could be the #1 result. But every search engine differs, and they handle different types of queries in different waysthat's one reason they have different results for the same query.

As you continue your education in search marketing, you will discover many articles that answer the basic question, "Just what are search engines looking for?" Each article has slightly different answers. Sometimes the articles contradict each other. Do not be concerned about that.

Search engines are fiercely complex, and they change all the time. In addition, "what search engines are looking for" strikes to the heart of a search engine's trade secretsits ranking algorithm. So, maybe the article's writer observed a few situations and concluded something that wasn't quite truethe search engines will never publicly divulge the truth. Or maybe what one writer wrote might have been true when it was written, but is not true anymore. Perhaps two writers performed tests with two different search engines, and what is true for one is not true for the other.

Unfortunately, divining how search works through observation is hopelessly subjective. You will read conflicting and erroneous information about what search engines are looking for, so you need to take everything you read with a grain of saltincluding what you read here. We do not have any inside information. We have lots of experience, just like most of those other writ ers, but what we write might not be any more accurate than anyone else's story. And by the time you read this, the search engines might have added a new wrinkle. If you believe you must keep up with the ever-changing algorithms, consult the resources in Chapter 16, "What's Next?"

But maybe you can take a different approach. What is more important than the details of how any particular search engine works is your philosophy for feeding them tasty spider food. How you think about what search engines want is more important than what they actually want at any particular moment in time. If your philosophy is to outsmart them at every turn, constantly tuning your pages to fit the latest ranking algorithm, that's one philosophyone we do not recommend. Wouldn't you rather use an approach that does not need to be changed every week, one that people without any special training can learn and stick to? We advise a philosophy of writing for your visitor firstyou will learn what that means next.