Purpose

• Internal structure of the site
• Maximize document relevancy for users
• Rule of thumb: good usability and accessibility practices

HTTP and URLs

HTTP 404 (Not Found)

No reason given as to why the request is refused

• Purge all 404s: either with 410 (Gone) or 301/302 (Redirect)
• Similarly remove all HTTP 501 (Not Implemented)

HTTP and URLs

HTTP 410 (Gone)

Resource is no longer at this location

Advises to the visitor to remove all references

HTTP 302 (Found)

Temporarily moved to a new location

Come back again here in the future. Maybe it is changed?

HTTP 301 (Moved Permanently)

On request, this resource will be found at a new location - trust me dude

Avoid duplicate (or very similar) content URLs. Point to an authoritative edition of the content

HTTP and URLs

• robots.txt
• Use htaccess/htpasswd for sensitive data or
• Server-side sessions to allow only authorized logins

HTTP and URLs

URL naming convention

• Machine URL --> human-friendly URL
• It should be simple enough to remember and type
• Cool URIs don't change!

HTTP and URLs

Server efficiency and code complexity

• Common downtimes, transfer speeds impact machine and human experience
• Consider the number of requests, page size, technologies in use...

HTTP and URLs

User-agents

Bots, screen readers, browsers, TVs, handhelds, projectors, ..

To customize what to serve

• Avoid using server/client side detection
• Impacts performance
• Unreliable

Separation of code layers

• Content (non-illustrative images, text.. anything that is "data")
• HTML: Semantics, Structure, Relationships
• CSS: Presentation
• JavaScript: Behaviour
• Flash: Interactive stuff?

Separation of code layers

• less interpretation increases accuracy
• minimise the time it takes to reach other URIs
• cover bare minimums and degrade gracefully
• don't hide information behind forms, logins etc.

Meta information

meta-meta-meta.. data

Data about the data

Used to provide supplemental information for the document

• description: around 150 characters ~ 1-2 sentences
• keywords: no longer relevant/necessary
• Machines use alternative methods to collect reliable words about the document
• Machines should not (?) trust humans

Meta information

<title>

Unique way to identify documents in a human readable format

• Around 80 characters ~ 1 sentence
• Be concise, honest (think usability) and unique
• Use Specific -> General keywords, e.g.
• Page name | Section name | Site title - Tagline

Content

• Context as opposed to content is king!
• concentrate towards two or three keyword combinations
• foo bar baz is a more focused search than foo bar
• indexing is in place to assist human searches
• keyword proximity

Document inter-connectivity

• Create a set of similar documents (tagging, taxonomy)

• Keyword proximity around links

  
  	funky		monkey	
  	  shWeeet!
  		ahoy hoy!
  foo			w00t
  	
  	
  	
  	
  	yo! I'm all the
  			way over here	
  

Content : Text : Link ratio

• Weight
• Density
• Location
• Saturation

Usability

• How can I get to where I'm thinking? ~3 steps
• Keep your document focused (break it down into multiple pages)
• Make use of @title, @alt attributes
• Text > Images > Flash (no ifs or buts)

External linking

Will someone think of the children users?

• Valuable resource => Create an external link
• User can find everything from you => You are the starting point!
• Happy user will link to your article!
• Place yourself in the middle of neighbourhoods (or build them)
• Stay away from the graph edges
• Links built the Internet!
• Don't encapsulate yourself from the rest of the internet!

Semantics and future compatibility

• Always write for humans! Let the machines do the catchup
• Text analysis (e.g., complexity, nationality, gender, age, tone, multiple-authors?)
• microformats ~ (s|S)emantic (w|W)eb tubes
• ... ?

References and Resources:

• W3C Link Checker
• HTTP/1.1: Status Code Definitions
• The Web Robots FAQ
• Apache Module mod_rewrite
• Cool URIs don't change
• Web Page Analyzer
• Linearizing Tables
• Ranking search results by reranking the results based on local inter-connectivity
• The Anatomy of a Large-Scale Hypertextual Web Search Engine
• W3C Markup Validation Service
• Microformats

Illustrations are from: Seybold (Licensed under CC 1.0)

Thanks!

• www.csarven.ca/presentations/internal-seo-guidelines-01
• www.csarven.ca/internal-seo-guidelines