Share

Types of web measurement

There are three prime methods for measuring internet usage,

• Site - website owners collect data on their sites through server logs or using a web analytics solution
• User - data is collected using software installed on the computers of panel users
  
• Network - data is collected from ISP networks

User
Using panel data web usage is estimated as usually global or local

• Alexa
• Compete
• Comscore
  
• Nielsen Netview

Network

• Hitwise

Site

Server logs

• Apache
• Redhat
• etc.

Free web analytics tools

• Google Analytics
• Yahoo IndexTools (not sure if this is free yet)
  

Enterprise web analytic tools

• WebTrends
• Omniture Site Catalyst
• Hitbox
  
• WebSideStory
  
• ClickTracks
  
• CoreMetrics
  
• SAS
  
• IBM Surfaid
  
• Unica Affinium NetInsight
  

Site measurement

Server logs

Server logs are daily files automatically created by a web server documenting every action that has been processed. If opened in a text editor these files would show raw data. A web analyzer application would be used to read and analyze the data into a format that can be understood, producing charts and graphs.

Typically the files contain the following information,

• Host provider
• IP address
• Operating system
• Browser
• Screen resolution
• Pages visited and times viewed
• Total time spent on site
• Files downloaded
• Cookie information (returning visitors)
• Origin of visitor (search engine/ search term used/ referral link)
• Ad click
• Pages bookmarked
• Search engine spider activity
• Error messages

Web Analytics tools

There are two methods of collecting data into a web analytics solution. Web (or on demand) using JavaScript tagging to site pages provides a more efficient method for data collection, plus allows for more information to be collected compared to server log files.

• Web (JavaScript tagging on pages)
• Software (analyze server log files)

Outline requirements and limitations

Management
To provide the overall requirements for implementing a web analytics solution. (also see Putting together a web analytics brief).

• Site objectives/ goals/ strategies
• Key performance indicators
• Reporting requirements and scheduling

IT
To provide details on the website's technical specifications. (also see Putting together a web analytics brief).

• Site architecture
• Servers
• Scripts
• pages
• applications/ objects/ etc.

Users

• Factor in the needs of all stake holders that would potentially use a web analytics solution
• Appoint a person to be a point of contact for administering the tool

Prioritize over analysis or reporting

What is more important, powerful reporting on KPIs (volumes/ counts/ ratios), or powerful analysis to assist actionable decisions?

Implementation

• Pages
• Ensure each page has a unique page definition (within the title tag)
• JavaScript tagging on all pages
  
• Ensure the tag is the last thing to load by placing it before the close body tag
• Best practice is to place the tag within an include file
• Ensure the tag is inline and not within a table or frame

• URL structures
• Ensure URLs are unique and easily interpreted
  
• Changing URLs could affect pages that have been indexed by search engines and search engine friendly formats

• Cookies
  
• Source, page and user attributes
  
• Inserting parameters into URLs
• Scripts - using small scripts to register cookies

• Objects and applications
  
• Tracking for rich media - Flash, Flex, RSS, videos, etc. (event tagging)

• Custom
  
• Conversion - Goal pages/ revenue values
• Conversion funnels
  
• Scenario analysis
• Campaign tracking

• Potential obstacles
• JavaScript url links (site overlays might be a problem and definitely no good for SEO)
  
• Redirects

• Validation
• Is the data being captured correctly?
  

Related websites

Good round-up of free and commercial web analytics tools:
http://www.conversionrater.com/2006/02/15/a-complete-guide-to-web-analytics-solutions/