The question came up again this morning. What’s wrong with the practices normally associated with User-Centered Design (UCD)?

Here’s a list.

• User research is horribly unreliable, either because it’s done poorly, not done to a great enough extent, or focused on the wrong people.
• Focusing on niche groups can mean alienating audiences that could otherwise use the product but weren’t part of the research.
• User research can be horribly expensive and time-consuming, and I have yet to work for or with a web company who was willing to devote an appropriate amount of time to it. Companies with industrial products might be different (Apple is a notable exception, of course), but on the web, things move fast. Since the ability to iterate is built into the web, it makes far more sense, financially, to iterate continually rather than put great effort into research prior to the start of a project.
• User research, as it’s typically done, results in a set of persona descriptions, which are, well, less than useful as project deliverables. Managers care about results. Numbers. They want to see progress, not fictitious character descriptions. They hired you to design, not write movie scripts.
• Far too many people seem to think you have to perform all new, application-specific research any time they start work on a new product.
• Reliance on UCD methods can lead managers to mistrust a designer’s instincts, and instincts are a huge part of design. I’ve seen managers on many occasions ask for all sorts of research-based validation on things that should not need it at all—the usability of a design pattern, the validity of a task flow decision, etc. It discounts the designer’s experience, skill, knowledge, and talent. It turns designers into scientists, and designers don’t make for very good scientists.

I could go on. (In fact, I wrote a series on an alternative for Peachpit.com.)

Even those who often advocate the practices included within UCD freely admit that, most of the time, these things are done wrong and done poorly.

To clear up a misconception, none of this means a non-UCD designer has no understanding of human behavior on the web, or that he blatantly discounts the importance of understanding it. I personally have sat through countless hours of usability tests, study psychology in my spare time, constantly engage in conversations with people about their computing experiences, etc. I couldn’t do my job without a great understanding of people. I just believe it’s far better to focus on activities rather than people.

Yes, this can mean talking to people who perform the activities you’re designing to support, but in many cases, you can become a Subject Matter Expert on an activity without ever talking to another person—particularly when designing web apps.

No method is perfect, but in my experience, UCD has far too many flaws to come even close. And considering there are so many great apps out there designed and built without an ounce of app-specific user research, it’s safe to say you can indeed create something wonderful by throwing traditional UCD methods entirely out the window.

Posted by Robert on June 24th, 2008