Last October I sat in on a presentation by Peter Sevcik at the NGN Conference. He discussed the problem of slow loading of web pages emphasized the need for web page designers to be aware of this important issue. Poor design often times contributes to a bad user experience. As a network guy I don't have a lot of influence in this area. But it did get me thinking.

It reminded me of the by-gone days when I was a part of the Corel Customer Advisory Board. Corel's VP of development (his name escapes me at the moment) expressed his concern regarding the performance of the applications in the Wordperfect suite. He gave each of his Product Managers a stopwatch and told them to get the response time's down. This was a good way to get measurable results. It also focused everyone on an important issue.

Peter Christy and Sevcik seem to be having a spirited debate about Internet application speed in the March Issue of Business Communications Review magazine. Sevcik categorizes network users into three zones: satisfied, tolerating and frustrated. He suggests that 10 seconds is at the upper end of the "satisfied" zone. Peter Christy disagrees. He feels that "the right goal is a fraction of a second."

It isn't my purpose to defend or challenge either of these views. It is more a matter of opinion, patience and circumstance. One day 10 seconds may be way too slow. At other times 20 seconds may not bother me. My preferences change often.

In my experience there are two main problems with access to web sites. These are:

1. Sites that are chronically unavailable.

2. Slow access.

The first can usually be traced to DNS. It drives me crazy to get complaints that are along the lines of, "my users are upset that they can never get to XYZ.com. This has been a problem for several months." These problems must be systematically addressed and corrected.

The second is equally upsetting but not as easily understood. To get a handle on "slow" access there must first be a baseline. I think that the users in the tolerating and frustrated zones are there because access speeds are appreciably worse than normal.

But what is normal? I don't know and don't pretend to have the answers. I do have ideas and opinions.

Here is what I think should be done. Find ways to measure the user experience in terms of web page response times. Take measurements to establish baselines. When the baseline measurements change find out why. Only then can we hope to address and satisfy the needs of those users in Sevcik's Tolerating and Frustrated zones.

Tomorrow: What all this means to UEN
11:16:44 PM    comment []