Website Architecture

January 29th, 2008

Executive Summary

Sharewell Group, Inc. SEM Team has assembled this Website Architecture Whitepaper as a result of extensive research conducted specifically on the affect of website file organization to search engine optimization. The result is the conclusion that having a well-organized and intuitive website architecture may help the various search engines categorize and rank pages within a website more effectively.

The scope of this whitepaper will include real examples of active websites that rank well in the search results with a limited number of inbound links. This report will also classify the core files that should be included on the root of the web server as to obtain the best results from a Search Engine Optimization campaign.

Implementation of this strategy is in no way a guarantee for higher ranking or traffic volumes. This document is to serve only as a “Best Practices” for website architecture.

The Common Mistake

2.1 Dump All Files on the Server Root

It’s not uncommon to find website owners who know very little about organization when it comes to building a presence on the web. The typical novice will create their website from either Microsoft FrontPage or worse, from a web interface such as those hosted by GeoCities.com.

These owners don’t know any better, so they neglect to classify the pages on the website and in turn are losing out on potential positions in the search results. Creating folders on a website is a way of categorizing content. Unfortunately, the quick and easy way to upload files and maintain simple relative URL’s is to simply neglect the categorization and simply dump (FTP) all files to the root of the web server.

2.2 Non-Architecture Mistakes

Page-level actions that are often overlooked by web designers that relate to the categorization of content mentioned above, include having a unique Title tag, Meta Description, Meta Keywords, and H1 heading tag appropriately added to every page of the website.

The context of how these items should be populated is out of scope of this document. However, they are significantly important to overall topic classification for the search engine spiders, and therefore should not be overlooked.

Sample Website Architecture

3.1 Explanation of Website Structure

According to SEO Expert Steve Wiideman, a website is like a catalog and should be organized as such so that both users and search engines can quickly and easily navigate. The architecture below may significantly improve a search engine’s crawl and index experience for this plastic surgeon.

Explanation of Website Structure

3.2 Explanation of 3.1 Landscape

The diagram in 3.1 above shows a recommended website architecture plan designed by SEO Expert Steve Wiideman. Though it looks simple enough, there are several different items that should be addressed.

There are no files on the web server that aren’t displayed to the user in one form or another. Development files (.psd, .fla, etc) are nowhere to be found. As with backups of a previous version of the website. A clean web server is easy to navigate and less apt to be hacked and linked to by a competitor. The last thing a site owner would want is to have an older version of a newer website continue to appear in the search results.

Keep a limited number of files on the root. All content can and should be categorized. There truly is no reasonable excuse to not keep the root of a website completely free from content, with the exception of a few verification files, sitemaps, and of course the homepage.

Directory nomenclature is everything. These directories should describe all the content within them and whenever possible be search terms that generate high volumes of searches throughout the various search engines. Example: /coolresorts/ is not a high volume search term, but /hotels/ however is a very high demand term.

Page nomenclature is just as important. The file name of the web page is displayed in the URL, which has been known to be part of the ranking criteria by several of the major search engines. See the example below:

Explanation of Landscape

3.3 Static Sitemap

Static sitemaps are pages dedicated specifically to internal link development. Text within these links act as an introduction to the page (for the search engines); thus these links need to be keyword-rich for maximum search engine optimization.

For sample sitemaps see the Link Analysis Report. The sitemap for the redesigned AndreAboolian.com website should be broken down by the directories above, and should include at least the top search terms outlined the Search Term Discovery. Each category should be represented by a heading tag (preferably H2-H3).

3.4 XML Sitemap

As detailed in the Link Analysis Report, this list of URLs allows the webmaster to quickly and easily notify many of the major search engines of new pages added to the website. Currently the two places to submit this file are:

Typically, each search engine will require a verification code be added to the root of the website. Get a Free XML sitemap for websites under 500 pages here:

http://www.xml-sitemaps.com/

3.5 robots.txt

The robots.txt file is a critical part of the site architecture. This file prevents the search engines from crawling specific directories and files; specifically, content that has little to do with the targeted search terms.

Sample directories to be excluding include /js/ and /css/.

Save time by building the robots.txt file using this tool:


http://www.mcanerin.com/EN/search-engine/robots-txt.asp

Bonus: Page Layout

4.1 Landscape of a Webpage Optimized for Search

» Global Navigation

The global (main) navigation should be built with Cascading Style Sheets (CSS) and be text-based so that it can be read by the search engines. Example: www.questinc.com/sitemap.html

Links included in the global navigation should be intuitive to the user and broken out by category (H3 tag) as described in the site architecture layout.

» Lower Navigation (Footer Links)

This lower navigation is typically less intuitive to the user and more understandable to the search engine. Rather than using terms such as “plastic surgery”, this navigation will house the most important search terms, such as “plastic surgery los angeles”. This navigation should NOT be hidden to the user (a Google terms violation), but should be discrete as to not disrupt the flow of the website.

» Footer Navigation

This simple navigation typically offers links to the website’s Privacy Policy, Sitemap, Contact Information and any applicable Terms and Conditions.

» Text / Paragraphs

All text should be presented in textual form and NOT embedded on an image. Vanity occasionally causes the site owner to post his/her name on each page of the website. The result is that the search engines start believing the page topic is the person’s name and the search term being optimized on each page.

» Heading Tags

Each page should have a heading tag (H1-H2), which can be formatted in the external stylesheet for readability. This tag should begin with the targeted search term and include additional text to not appear as a spam technique (called keyword prominence).

» ALT Attribute on Images

Each image should house an ALT attribute (Alternative Text) for the visually impaired. Many search engines will give a site additional credit for accessibility and favor search terms being included tactfully within this attribute.

» Social Buttons / RSS

None of the competitors of www.AndreAboolian.com offer Social Bookmark buttons, which allow users to quickly add inbound links to their favorite social network. Really Simple Syndication (RSS) allows external websites to pull updated content automatically (including links).

Additional comments

5.1 Content Management Systems (CMS) Requirements

Many CMS applications are built for usability and not necessarily for Search Engine Optimization. The following is a list of criteria that a CMS must have in order to be considered “optimized for search engines”:

1. Sections and pages within the CMS must return static URLs using user-defined nomenclatures. Example:

  • Good: SomeWebsite.com/travel/europe/spain.html
  • Bad: SomeWebsite.com/?sec=1243&sub=4532&pID=100056

2. The CMS must be able to auto-generate the static sitemap page at the root of the website using user-defined link text OR using the HTML title tag.

3. The CMS WYSWYG editor should include custom fields for the user to enter their own:

  • Page Title
  • Page Description (HTML meta description)
  • Page Heading (HTML H1 tag)
  • Image ALT text, as applicable
  • Link Text (see 5.1 Item 2 above)

4. The CMS should integrate efficiently with commerce and syndication applications, which may include the use of a subdomain (shop.somewebsite.com) and data feed pushing (Google Product Search, Shopzilla, etc).

5. The CMS must not render unneeded source code to the browser, such as JavaScripts, CSS styles, Java, etc. These scripts should be in an appropriate folder or embedded into an AJAX module that is not displayed in the source code of the rendered website.

6. The template used within the CMS must be W3C compliant (CSS & XHTML)

7. The CMS must not render 2 or more copies of the same content. Wordpress for example, renders three copies, one on the homepage, one in Archives » Month, and one as it’s own isolated page on the website (myblog.com/date/my-blog-post.html).

Custom Code & Commerce

Specifically for static websites extending to commerce applications or offering non-keyword-rich content, a subfolder or subdomain should be used to isolate this content from the rest of the website. i.e. shop.mystore.com, mystore.com/shop/, members.mycommunity.com, etc.