Summary

Web Analytics Tutorial

 

Lesson 6 – Design Considerations

IN THIS LESSON
* Browsers
   Operating Systems and Platforms
*  Browser Features
   Screen Real Estate and Colors
* Language and Speed
   Language
   Load Speed

Browsers

Figure 1. Sample Web Graphics
Figure 1. Graphics and text can
render differently on browsers.
A major element of the Web’s popularity is the cross-platform ability of HTML. A web site can be run on one type of computer and all other types of computers can access and display the content. However, HTML is a markup language, not a layout language. Using it for layout, combining text with graphics, and trying to include other technologies, can be a complicated mess if you have to verify that the site looks perfect on all browsers, on all platforms. Figure 1 shows an example of a company using graphics for their logo at the top of a page, then a line of text links below for a menu of content sections. Notice how the intended page would have all the menu items on one line, but because the window width was too narrow, the phrase ‘About Us’ was wrapped to a second line (in this contrived example). Problems like this are unfortunately quite common on the web. This could be related to font settings, window settings or just subtle differences in the way different platforms render items from a web page.

Generally it is not time-efficient to design for every available platform and browser. There are several hundred web browsers and over one hundred different operating systems. In most design projects you can accomplish 80% of the work in 20% of the time. This is often referred to as the 80/20 rule. The break in value of time can be more or less than 80% depending on the project, but the same principle applies – at some point it takes too much effort to design further. This is true in web site design as anywhere else. You can spend lots of time testing your design on each available platform, but if you know what browsers your visitors have, then you can focus on designing for the most popular and make your work more efficient.

Figure 2. Browser Report
Figure 2. The Browser Report tells you the percent of visits
for each browser so you can design for the majority of visitors.
The Browser Report in Figure 2 shows the top few browsers (or ‘user agents’) used to visit the site. The top entries are usually versions of Microsoft Internet Explorer or Netscape Navigator. Thereafter you may see some more obscure browsers and the balance will usually be robots (which you can inspect in detail in the Known Robots report.) By adding up the values in the % of Visits column, you can determine which browsers comprise the top 80% and choose just to test for those. Alternately, you may see an obvious break in the values which makes more sense to limit at. Or you might select to test with all major browser versions listed in the first page. Whatever your rule, you can increase the efficiency of your design testing by knowing which browsers your visitors use and testing for those. As time passes, new browsers come out, and popularity changes, so you should regularly monitor this report (and other design reports) and change your site as needed to keep in step with your visitors.

Operating Systems and Platforms

Figure 3. Platform Report
Figure 3. The Platform Report gives
insight to visitors’ operating systems.
Unfortunately, a given browser version may not even render the same on different platforms. Microsoft Internet Explorer 5.1 on Macintosh OS 9 lays out pages differently than the same version on Windows 2000. So it is also important to know which platforms your users have. Figure 3 shows an excerpt of the Platform Report, which shows what operating systems visitors have. For operating systems that can run on different hardware, the CPU architecture is given, for example, ‘Linux x86’ or ‘Macintosh PPC.’ Again, depending on your design rule, you can effectively choose the most efficient platforms to focus on when testing.

As you are not usually concerned about robots when designing, the Platform Report aggregates all known robots into a single listing, ‘(Known Robots)’. It also aggregates all user agents that do not include operating system information into the listing ‘(Unknown)’. These are usually also robots of some kind, so you can ignore these too when choosing platforms for testing.



Table of Contents | 1: What is Web Analytics? | 2: Where are My Visitors Coming From? | 3: Search Engines | 4: Advertising | 5: Revenue Modeling | 6: Design Considerations | 7: Determining Visitor Behavior Patterns | 8: Examining Subsets of Traffic  | 9: Incorporating Business Goals | 10: Bandwidth Management | 11: Site and Server Diagnostics | 12: Investigating Troublemakers | Appendix A: Making Reports More Usable | Appendix B: Technical Details of Metric Accuracy

Copyright 2002 by Summary.Net - Updated 16.Apr.2002