Academic Affairs Web Alliance

Search engine technology has changed greatly in the past few years. Previous methods to acquire and maintain a top search ranking have taken a back seat in the algorithms of the major players (Google, Yahoo, Ask, MSN Live). Up-to-date, relevant content with strong in-bound links are the “new” Search Engine Optimization (SEO) basics. Additionally, the proper usage of HTML, CSS, and Javascript will allow search engines (and your users) to more effectively find the content they desire.

To Do

Acquire strong inbound links from reputable sites

Arguably one of the most important ways to increase a site’s ranking is to acquire strong inbound links from reputable sites. Each link on the World Wide Web counts as a “vote” for the destination page and website. The greater the “reputation” of the referring website, the more weight that is given to the vote. For example, a link from CNN to a page at USF will probably count more than a link from a local news outlet. Additionally, the link’s weight is affected by how many other links are on the page. If Site A has a link to Site B on a page with 100 other links, that particular link’s vote will count less than if there were only 5 other links on the page. This approach helps reduce the effectiveness of “link-farms” that contain hundreds or thousands of links but no actual content.

That being said, link to your website from as many places as is reasonably possible. Include relevant URLs in media releases and press statements (be sure to use relevant anchor text…see below). Include a link in the signature section of e-mail. Include clickable links in PDFs that are published to the web (with a few caveats, PDFs are generally archived by search engines).

Use relevant anchor text

Links to a page or website should always be inline and anchored with relevant text. In the following examples, #1 will most likely be more effective (for searchability and readability) than #2 or #3:

1. Recently, USF received classification as a Research University (Very High Research Activity).
2. Recently, USF received classification as a Research University (Very High Research Activity). See http://www.carnegiefoundation.org/classifications/sub.asp?key=748&subkey=13889&start=782 for more information.
3. Recently, USF received classification as a Research University (Very High Research Activity). For more information, click here.

Use relevant keywords in the URL’s domain/folder/filename

A portion of the ranking for a page or site is derived from the domain name, the folder name(s), and the filename. When creating a page on the topic of our “faculty authored textbook policy,” Item #1 in the below list is probably a better name/address than #2 or #3.

1. http://www.acad.usf.edu/faculty/policy/faculty-authored-textbook.asp
2. http://www.acad.usf.edu/site/fac-txtbook.asp
3. http://www.acad.usf.edu/index.asp?page=faculty+authored+textbook

Option #3 is a poor choice for readability’s sake, but the link and content will probably be indexed and ranked even though it uses a querystring (the portion of the URL after the ?). There is significant discussion on whether or not querystrings hurt search engine rankings. It appears that, in the past, querystrings were detrimental, but many engines can now index and rank the content. However, there are still some who argue that if all else is equal, querystrings will rank worse than non-querystring URLs. Matt Cutts has this to say in the comments section of a blog post:

…dynamic pages can receive PageRank just like static pages can. I’d avoid session IDs though.

It is a common misunderstanding that “dynamic” content will hurt the ranking of a page or site. This is not the case. Web pages can use any type of dynamic processing (as long as your server supports it), whether it is .ASP, .ASPX, .CFM, .PHP, etc without fear of a penalization in search engine ranking. However, there does not seem to be a final word on querystrings yet. If querystrings are required, be sure to use human-readable text in the string.

Main site navigation: use list-based menus, not JavaScript arrays or Flash

Definitive answers are somewhat elusive, but most research indicates that search engines do not completely and properly parse Javascript (JS) or Flash. Compounding the situation further are screen readers and hand-helds that also do not parse JS. For many reasons, the best approach is to use a list-based menu augmented with JavaScript to make the menu interactive. Properly done, the menu will “degrade gracefully” for devices that do not support JS.

Generate a machine-readable sitemap

Humans often revert to a site map (if provided) when they cannot find targeted content on a website. A new standard for search engine spiders has recently been developed that builds on this idea. Sitemaps.org is the product of a partnership between Google, Yahoo, and Microsoft. The main goal is to create a machine-readable map of a website that’s found in a standard place, similar to the approach taken by robots.txt.

Provide a robots.txt file

A robots.txt file tells search engines which directories they can and cannot search. Most, but not all, spiders will follow the rules specified within the file.

Use 301 redirects, not meta refresh or Javascript

Pages in a website that have been moved or relocated should direct all requests for the old page to the new location. There are a number of ways to accomplish this, but webmasters should virtually never use meta refresh to specify where a new page should be found. Instead, if the web server can parse PHP or ASP, use the following technique:

• Open the old page in your web editor using. Be sure you’re in “code view” mode.
• Remove all code and content from the page. The web editor should now show a completely blank page.
• Insert the appropriate code from below based on your webserver’s capabilities

PHP

<?php
Header( "HTTP/1.1 301 Moved Permanently" );
Header( "Location: http://www.new-url.com/folder/file.htm");
?>

ASP

<%
Response.Status="301 moved permanently"
Response.AddHeader "Location", "http://www.new-url.com/folder/file.htm"
%>


• Replace http://www.new-url.com/folder/file.htm with the address of the new page
• Save the file
• Visit the old address in a web browser
• You should seamlessly be redirected to the new page, with the new address appearing in the browser bar

There are a few situations where the above solution might not work:

• the server cannot run PHP or ASP
• the file ends in .HTM or .HTML and those files can’t be mapped to the ASP/PHP engine

in which case the alternative is to use the “redirect” directive in Apache or IIS.

Generate unique, proper meta and title tags for all pages

Although meta keyword tags are no longer the main focus of search engine rankings, they are still valuable in the overall ranking scheme. Title tags are also important. Be sure to include unique, relevant words and phrases for each page in the site. In other words, every page within a site should not use the same, unmodified set of meta or title tags. Search terms that are important for a site should appear in 3 places: the body of the page, the title, and the meta keywords section. However, be sure not to “pad” the site with words that aren’t displayed or repeat terms over and over, even in the meta tags. Most search engines can detect this and will penalize the site ranking. A general guide is to use a total of 11 words (this includes phrases) in the meta keywords section without repeating any one word more than 3 times.

Choose the proper website address, e.g. www.usf.edu or usf.edu, but not both

Website addresses may contain a “www” preceding the domain (e.g. www.usf.edu or www.acad.usf.edu) or not (e.g. usf.edu or acad.usf.edu). In general, either method is acceptable as long as all inbound links and internal site references use the same address. This is referred to as a “canonical” URL issue. Because search engines consider usf.edu and www.usf.edu, and usfweb2.usf.edu as completely different websites, it is imperative to address this issue. Matt Cutts, a Google engineer, has 3 suggestions which should be followed.

Use tools supplied by the search engines

Google’s Webmaster Tools are a great place to start. A webmaster can

• create a sitemap.xml file
• view indexing problems with their website
• resolve canonical issues (www vs non-www)
• view inbound links
• and more

Yahoo Site Explorer offers a few similar features, allowing users to explore a website/URL to resolve indexing issues. Generic but helpful resources include MSN Live’s Site Owner Help and Yahoo’s Search Basics Help.

Use dashes, not underscores in URLs

Using a dash in a URL is not only more readable for a human: search engines prefer it too. For additional reading on why, a few links are below:

• Matt Cutts, Google engineer, Dashes vs. underscores
• Vanessa Fox, Google Product Manager, Vanessa Fox on Organic Site Review session
• Underscore treated as a “word” character by regex

In the recent past, there seems to have been a large shift toward the dash rather than the underscore by a number of major websites, e.g.
http://tech.netscape.com/story/2007/04/10/yahoo-viacom-strike-search-deal. Surprisingly enough, CNN and some other major players are often using periods as delimiters, such as
http://www.cnn.com/2007/TECH/biztech/03/26/tvguide.video.search.ap/index.html
The period is an interesting approach but no one is certain how it will play out. Traditionally, periods have been reserved for the subdomain/domain and the page extension (e.g. .htm, .asp, etc). Time will tell how search engines treat the approach.

At the current moment, all recommendations point to using a dash as a delimiter in URLs over any other character.

Conclusion

Many of the approaches discussed above can be implemented on a website in a short period of time. Others, however, are difficult to employ on a pre-existing site. Going forward, a webmaster should consider using all of the above approaches on every new page that is created within an existing website. As time permits, old pages can be updated to reflect the new techniques required to receive top search engine rankings.

Comments, suggestions, requests for assistance

This site was created to foster discussion so please leave comments, suggestions, or corrections in the Reply section below. No registration is required (but feel free to do so if you like!) and your e-mail address will never be published. In fact, you are welcome to leave a fake address such as no@thanks.com if you like.

For a more in-depth discussion on ways to utilize the above tips on your site, feel free to contact me at tthompson@acad.usf.edu or 813-974-1630.