|
|  |
User Experience (Usability)
322 hits
| Oct 04, 2003 8:29 pm |
|
re: re: re: Winning the small battles... |
Chris Hubbard
 | |
My approach has always been a bit different. Janna is absolutely correct, arguing with a client is like wrestling with a pig in mud. After a while you realize they they like it, and you're all muddy. So what I've been doing is asking the client to set objective usability goals before we roll up our sleeves. I usually start one of the meetings with a question like "What percentage of users need to be able to successfully complete the top 5 tasks on your site without requesting help?" Typically the first answer I get is 100%, and for a site with nothing going on, that *might* be obtainable. However with a complex site (like e-commerce, or networking, or...) typically I've seen no more than 60% - 70% of the given users can complete the top 5 tasks in NN minutes without help. I've seen web sites (a lot of them actually) that start out about 5%- 10%, and after a round or two of usability based change, end up being 15%-20% usable.
The numbers need to speak for themselves. If the client targets 70% and after a round of testing end up at 35% then they need to come back to you for how to improve the site. It might turn out that the flash based splash page has no impact on the overal usability, it might turn out that the 150 animated gifs actually improve usability(?!?). The point is that it doesn't matter what the client thinks, or the client's nephew/neice, or neighbor's cousin. After a clean, well-thought out usability test, there's little room for interpretation. The site either does or doesn't meet the goals.
Measuring the usability of a site is a simple process.
* Determine what the top 5 tasks are for the site.
* Determine how long it should take the average/mean user to complete each task. (a rule of thumb, take how long it takes you to do it and multiply by 2)
* Determine how you the tester can tell that the user has completed each task successfully.
* Write down what it is that the user needs to be able to accomplish, have about 10 different people read it out loud (very important) to you. When they're reading it out load to you, listen for where they pause or stumble or take a breath. If it's in the middle of a sentence, rewrite it.
* Write down what you are going to tell the testers. Write everything down. Everything.
* Find 10 average users, read from your notes what you need start the test, start the timer, hand them the instructions and watch/video tape them. Part of the instructions is for them to tell you when they've succeeded. You don't tell them, they tell you. When they tell you they've succeeded, or the timer runs out, say "thank you" and go on to the next task.
* At the end of all this you should have notes for testing 50 tasks (10 users, 5 tasks each, 10 x 5).
* Collate and average. Let's assume we have the following numbers:
Task A: 5 succeeded, 5 failed = 50% usable
Task B: 7 succeeded, 3 failed = 70% usable
Task C: 2 succeeded, 8 failed = 20% usable
Task D: 8 succeeded, 2 failed = 80% usable
Task E: 3 succeeded, 7 failed = 30% usable
Average usability for this test, for all users, for all tasks is 50%.
Now it may turn out that the users doing Task C got caught up watching the flash splash page and so ran out of time. At that point perhaps one of the recommendations would be to remove the splash page.
Heh. That's rather more than I set out to write.
Chris
Private Reply to Chris Hubbard (new win) |
|
