Humboldt Road Trip

Well, another year and another Alumni Reunion up in Humboldt. I have more pictures, but here are the first batch from the return trip. We stopped at the Hopland Brewery in Hopland, CA for lunch and more drinks. Our posse consisted of Scott (Pinto), Justice (Aardvark), Liz, Rhonda, and me. This place is one of our favorite watering holes on the way up to HSU dating back to our college days.

Humboldt Really Got Rolled Over

Well, I survived my vacation in Humboldt this weekend. My Alumni reunion was absolutely insane...9 hours sleep over 3.5 days, reconnected with brothers I haven't seen in a decade, and did a brewery tour of the Humboldt area...sampled all the beers I grew up on back in the college days. It was an experience. Photos will be posted later on after I get some more sleep.

Business Metrics Justify User Experience

Macromedia (pop-up) recently redesigned their web site using internal staff. What's interesting about this change in the site is not so much the flash heavy design (which I am not a fan of), but the fact that they took user feedback into account and then published the results of the changes with business metrics.

Some of the key areas they list for business metrics include:

• store conversion
• membership registrations
• Exchange downloads

As Mark from Good Experience says....

...user experience work is worthless without the business metrics. Why improve a task if it doesn't also improve the company, or organization, or some larger context? Usability for its own sake, without some larger good, is nonsense. It's like going on a road trip - with the car up on blocks. Or hiring the best chef - to cook up some gravel for dinner. It's going through the motions without some useful result.

This is just another example of grounding user experience development processes with measurable metrics for change before we even begin the process.

Challenging Usability Myths

Back in August Janice Fraser of Adaptive Path voiced support of Jacob Nielsen's claim that you really only need 5 participants to get useful usability results

"Instead of bringing in a large number of test subjects, keep the scope small. I often disagree with usability guru Jakob Nielsen, but he's dead-on when he says that you only need to test with five people to obtain good results. Any more than that yields diminishing returns, so it's usually a waste of time and money."

The problem with this approach, as was just posted on Jared Spool's list serve, is that the "5 person rule" was devloped to test small software applications and not large, complex, multi-user group web applications that are the norm today on the web. In fact, the author claims that Nielsen's paper from 1997 did not specifically specify that 5 users was enough for all user testing.

What's intersting about this contrast in reality is that many decisions about user testing from small to large web applications have been based upon the assumption that 5 users is always enough to get a good sample. Unfortunately this is not really true, and we are not necessarily getting an accurate picture of the usability of a large web application.