Posted in Website Usability Testing	December 21, 2006

Experts on Website Design and Website Usability Testing Can Disagree

By Don Nicholas, Editor Mequoda Daily and Library

"Never redesign your website," says one of my esteemed mentors and colleagues.

Disturbing words, especially because, as a website design expert and information architect, I spend 50+ hours a week explaining to clients why and how they must redesign their websites.

"You're wrong about that," says another colleague, relieving my discomfort and validating my position.

Here's how it happened.

The bad boy of website usability testing

At the recent SIPA conference in Las Vegas, I was both surprised and thrilled to see a number of online publishing experts appearing as guest speakers.

Among others, I was especially eager to hear from two of the real industry luminaries:

• Jared Spool, CEO & Founding Principal of User Interface Engineering, in my estimation, is one of the top website usability experts in the world.
• Dana Todd, Executive Vice President of SiteLab, a leading search engine optimization consulting firm, is also the current president of SEMPO, the Search Engine Marketing Professional Organization, of which I am a member. I don't know her personally, although I've heard nothing but good things.

So my anticipation for their presentations increased.

Now, Jared likes to characterize himself as "the bad boy of website usability testing." His keynote address included a review of a case study in which a 25-year-old male could not complete a purchase on the Gap.com website.

Jared's point: Anytime a user gets to a friction point in task completion at a website—because there is no sales person, no toll-free number, and no online chat present—and even if it is present they may not be willing to use it—he stops. He can't complete the task.

In this instance, the 25-year-old was buying a sweater for his girlfriend but the Gap site was displaying sizes for women's clothing in a different way than he understood them. He knew she was a size 6, but the sweaters on the site were sized XS-XL, with no comparison chart stating what a size 6 would be in a sweater. This design flaw stopped the user cold in his tracks when he was trying to make a purchase.

Bad website design means no task completion means no sale.

Only through website usability testing could you discover that what you might otherwise consider to be a perfectly serviceable website, in fact has key information missing for a large portion of its audience.

The monetary value of task completion

For those of us who sell information products online, task completion rates translate directly into conversion rates and higher revenue per thousand impressions, or higher revenue per thousand emails sent.

If you're an advertising-driven publisher, task completion rates translate into more time spent online, higher user satisfaction, greater customer loyalty, and more ad impressions, which also enable you to generate more revenue.

The purpose of iterative testing

Jared Spool explained that both Yahoo and Amazon are iterative design testers. That is, they do frequent website usability testing to determine what might be wrong with their websites. They demonstrate tremendous discipline by envisioning solutions, mocking up these proposed solutions, and then doing additional website usability testing on the mock-ups.

Iterative testers like these giant online publishers often do multi-variant or A/B/C split testing to verify that their proposed solutions will, indeed, increase task completion rates.

And then, my heart skipped a beat

"In fact," based on what we've seen from Yahoo, Amazon and eBay, you should never redesign your whole website."

Here was one of the world's most renowned website design/website usability experts saying something that contradicts what I do for a living. I had to stop for a moment and realize that I disagreed with one of my mentors—one of the people I hold in high esteem.

This was deeply disturbing—that Jared Spool was telling a group of my peers that they should never redesign their websites, while I regularly tell them that their websites are so bad, that they have no choice but to redesign them.

I thought about this over the next several hours and came to the conclusion that Jared was wrong. Because my clients are not Yahoo, Amazon and eBay, who have already spent millions of dollars on website architecture.

Often, the reason a website redesign is necessary is that the taxonomy, lexicon, and language—the information architecture of the website—is simply wrong. Specifically, the language that enables (or hinders) the user's ability to navigate around the website is wrong.

It's wrong because it's not the way the user thinks. It's wrong because it doesn't reflect the product portfolio that the publisher offers.

We can validate this wrongness by website usability testing and search engine keyword research, or as Dana Todd calls it, "findability."

The need for "findability"

So imagine my glee when about half way through her presentation on Friday morning, Dana says, "Jared Spool is wrong. Websites often need to be redesigned."

If your information architecture is not right, if you're not findable, if your website does not dovetail into the Internet with some level of ease, then you have no choice but to start over, she said.

At Mequoda, we don't design single websites. They may sit on a single URL, but we design clusters of websites, grouped together in what we call a web network. That enables us to not have to redesign a website after we've gotten the taxonomy right.

Jared is wrong. In most cases, saying "never" is a bad idea.

At the same time, Jared is right. If you get the website architecture and taxonomy right, then you may never again need to redesign your entire website again, except to make continuous small improvements based on iterative design. Your website will slowly improve and the user will barely notice the changes as they will happen slowly and constantly.

But often times, a website that is four or five years old—that was constructed by a graphic artist with no training in search engine marketing, information architecture or direct marketing—will need to be completely redesigned.

At the end of the day, you may need to redesign your entire website, just as you may have to knock down a house that is too rickety to be remodeled.

And so I left the SIPA conference feeling validated.

References

SIPA—the Specialized Information Publishers Association

User Interface Engineering

Jared Spool

SiteLab

Dana Todd

Search Engine Marketing Professional Organization - SEMPO

Gap.com

Iterative design

Recent Mequoda Website Usability Testing Tips

How Print Magazines are Losing Core Readers
posted in Website Usability Testing | September 15, 2006

Eight Website Usability Terms you Should Become Familiar With
posted in Website Usability Testing | August 24, 2006

Using Card Sorting to Understand How Users Organize Content into Clusters
posted in Website Usability Testing | August 22, 2006

Tips for Deciding Which Words to Use on your Website
posted in Website Usability Testing | August 21, 2006

Three Reasons to Use Paper Prototyping Before Building Websites
posted in Website Usability Testing | July 6, 2006