HTML 4 Articles

This page is for articles of interest to the Web programming community. Feel inspired? Let me know what's on your mind and maybe I'll publish it here. Need more explanation or background about a particular topic, let me know and I'll try to use that topic in a future article.

Web Statistics (Part 2)

After I published the first part of this article in July, 2000, many readers wrote with suggestions for sites with browser statistics. Thank you all for your help. If you're trying to figure out what you should be using (and not using) on your Web pages, you've come to the right place.

First, many readers pointed out that while www.statmark.com doesn't work, http://www.statmarket.com does.

Greg Wenden offered a link to the quite interesting Browser News site.

James Belleau mentions that some interesting data can be found at Cyberatlas.

Kris Driessan suggested a look at Superstats if you want to analyze traffic to your own site. I'd also mention Darryl Burgdorf's Weblogs for analyzing your own site traffic. Many ISPs offer a free (already-installed) service as well.

Larry Pierce found an interesting site that not only provides statistics but also does some thinking about the numbers: Web Design Guide.

Finally, Kirk Carter at Service First Webmasters, Inc had some very helpful pointers. He writes:

Because of the large numbers involved, the greatest error in the Hitbox demo numbers will be in the extremely small categories, like Lynx or Opera users. Who designs a separate version of every page for Lynx? For practical purposes, then, the least accurate statistics are not important. I design for 95% of visitors because accommodating the final 5% is too expensive for my clients. The bottom 5% are not as likely to exhibit the desired demographic anyway.
If you use these numbers, be sure to click on the Week or Month tab in each category to obtain the broadest sample. The Year tab gives even bigger numbers, but usually we are interested in what's true now, and browser share (all the stats, really) change too fast to make that an accurate measure. Variations in Ezskins site promotion over that year will also have an unknown effect.

One more thing: By comparing two or three sample months in the past, you can get an idea of how fast, say, the decline of 640x480 is occurring. Projecting into the future, you can set a reminder in your calendar for re-evaluating your design criteria.

Thanks to everyone who contributed!

Web Statistics (Part 2)

Archives