Designing by Numbers Statistics and No Lies!

Designing by Numbers Statistics and No Lies!
By Pamela Heywood

Stats tell us a myriad of things, some even useful and we're accustomed to using them in our marketing and selling processes. Frequently abused and made to say what WE want, but here I am merely considering them in ways that influence site design ... and, I believe, letting the numbers speak for themselves.

Before finalising my own choices for the recent (and always in progress) re-design at TuCats Design I had a look at my stats and at various global statistical information online.

For this exercise, I took a glance at the overall picture via the third-party service I use from TheCounter.com -- all my pages have their invisible option installed. Raw log stats will tell you a whole lot more, but this serves the purpose.

OK, I'm lazy and I like getting their weekly report by email.

The information (percentages) for my site are those used for the analysis below. The global stats I consulted didn't show any fundamental differences from these findings, i.e. they should apply to you too unless your site only caters to some specific group that is far and away from the average.

MONITOR RESOLUTION STATS:

* Unknown (2%)
* 640x480 (9%)
* 800x600 (61%)
* 1024x768 (23%)
* 1152x864 (1%)
* 1280x1024 (1%)
* 1600x1200 (0%)

It wasn't that long ago that there was a case for making sites at a fixed 600 pixel width to fit the 640 x 480 resolution. For single-product sales letters, I think there is still a case. It looks better than having lines stretch across too wide a portion of the screen, even if the visitor does have 1024 x 768.

If you have more content to link to, you need space for menus and stuff. Is it safe to use a bit more screen real estate? I take my lead from Boogie Jack - http://boogiejack.com -- in that when a group drops below 10%, then it is safe to stop labouring too hard to make things absolutely and perfectly compatible.

The percentages above too are taken from cumulative data -- that is those built over time. What they don't show is that the 9% using 640 x 480 could have visited months ago. I certainly know that percentage has been dropping as time has passed.

So, to fit in with the now most common 800 x 600 resolution, I am using a fixed width of 750 pixels for the tables that form the basis for my design. Now it's true that you could just design in 100% widths so that it will adapt to everything, but it is just so much harder to do, especially if you want to use tables with multiple columns. The chances are high that you'll put some image somewhere which mucks it up. Been there!

As you can see, 1024 x 768 has now jumped up into second place (I use that resolution myself) at a pretty meaningful 23%. Yet I see sites every day that look absolutely awful at that resolution because they were designed for smaller sizes, but using 100% width tables. You really need to consider this group now.

However, it is easy to test and easy to change your resolution. For the vast majority of you using Windows 98. Simply Right click on any blank portion of your desktop and select "Properties" from the popup menu. From there, select the Settings tab.

You'll see a slider control in the bottom-right of that window, which will probably go from 640x480 to 800x600 and 1024x768 on the right. Choose any option and you'll have 15 seconds to "test" after which the machine will restore automatically to your previous settings, unless you tell it otherwise.

Another way is to get SilverThingy from Sausage Software "Test Web pages in a number of screen configurations. Using either Internet Explorer or Netscape you are able to preview your Web pages in the three popular screen resolutions (640 by 480, 800 by 600, 1024 by 768). This allows you to make certain your Web pages will look great on everyone's screens regardless of their settings!"

BROWSER USAGE STATS:

* MSIE 5.x (70%)
* Netscape 4.x (14%)
* MSIE 4.x (9%)
* Netscape comp. (2%)
* MSIE 2.x (1%)
* Opera x.x (0%) * MSIE 3.x (0%)
* Unknown (0%)
* Netscape 3.x (0%)
* Mozilla 5.x (0%)
* Netscape 2.x (0%)
* Netscape 1.x (0%)

The days of having to be absolutely pedantic about "cross-browser compatibility" are gone.

With a total of 79% of visitors able to see most or all of the extra goodies that you can do, particularly with CSS, in IE 4 & 5, use 'em. Just take care that they are only "lost" on Netscape users and don't turn out looking funny.

Forms are my bug-bear. You can format neat little boxes to fit in small spaces in IE, then they take up half the page in Netscape. I still have some to fix. If you design for MS IE, I'd advise you, from bitter experience, to take another look at yours.

Keep a copy of Netscape on your machine for testing as I do, or save heaps of HD space and pop along to AnyBrowser.com -- "Your Source for Browser Compatibility Verification".

OPERATING SYSTEM STATS:

* Win 98 (69%)
* Win 95 (15%)
* Win NT (6%)
* Win 2000 (2%)
* Unknown (2%) * Mac (1%)
* WebTV (1%)
* Unix (0%)
* Win 3.x (0%)
* Linux (0%)
* Amiga (0%)
* OS/2 (0%)

Nothing here we didn't know already! With only 1% of people using a Mac, you really have to make a decision. Do you really need to labour creating alternatives such as .pdf flavour eBooks or do you cater to the vast majority and stick with .exe? Your call!

JAVA/JAVASCRIPT USAGE STATS:

* No JavaScript: (5%)
* JavaScript 1.2: (1%)
* JavaScript 1.2+: (94%)
* Java enable: (92%)
* Java disable: (2%)
* Java unknown: (6%)

All you can conclude from this is that it is safe to use both these technologies. Whether you should or not is another matter!

If something can be done with CGI (drop menus, for example), use that in preference to JavaScript. Similarly, use the SSI Echo Directive for preference on things like dates.

If you MUST use Java, make sure you have a water-tight excuse and that you NEVER use it on your front page. I just hate having my browser crash the moment I enter your site or waiting a week for your "bells and whistles" navigation to load before I can go anywhere. How much of your site will I surf? Yep, you guessed it!

The conclusion from these stats is that the "average visitor" -- ignoring for one moment that the average visitor is actually a real live HUMAN BEING -- uses Microsoft Internet Explorer 5.x with a resolution of 800 x 600, running on Windows 98, with Java and JavaScript enabled.

If you design with this Mr/Ms. Average in mind, but just taking care not to completely upset any minority group, you'll come up with a happy-medium that makes for a good browsing experience for them and a heck of a lot less stress for you.

We could all use things that make life easier and that's no lie!

About The Author
Pamela Heywood is webmistress of TuCats Design - Building Your Online Business Instinctively. Subscribe to the weekly TuCats Mewsletter (sic) and get regular FREE hints, tips, articles and resources.