Are you about to take on a big website redesign? How about rebuilding that clunky-but-critical software application? Before you dive in, remember that the final arbiter of quality is not you, it's your users. Here are a few steps to better understanding their needs and behaviors before you spend any precious programming dollars:
Do your user research
Start with any quantitative data, such as analytics, that you already have to see what your users are (or aren't) doing. For additional insight, you can user-test the current site or software to see firsthand what delights and what frustrates your users. Talk with colleagues in sales or customer service to learn current and persistent user issues. Even if this research data already exists in a report somewhere, make the time to talk. The empathy engendered from an actual conversation with people in the trenches will naturally equip you to make more user-centered design and development decisions.
Build a prototype
Actually, make that prototypes (plural)? no one creates a perfect prototype on the first try. But that's the idea: to fail as quickly, as cheaply, and as often as possible knowing that each iteration gets you closer to a solution worth building. Certainly you can build effective prototypes with HTML or Flash, but Acrobat, Powerpoint, and even paper and pencil are still excellent tools to get your ideas into a tangible format. In doing so, you can better communicate, evaluate, and test your ideas. Speaking of testing?
When some think of user testing, they imagine white lab coats and clipboards. Unfortunately, many also imagine delays and extra expenses. When forced to choose between this and no user testing at all, most choose the later. For shame! On smaller projects or those with a wicked-tight deadline, take the guerilla approach: find 6 to 10 coworkers, parents, spouses, neighbors (whoever is willing to help) and observe them individually as they complete one or two of the most important tasks on your prototype. This won't give you all of the insight or fancy reports that formal usability testing provides, but testing even just one person is 100% better than testing no one. The results might surprise or even frustrate you, but better to know these things now than after the project is otherwise done.
The right design
It's true that we human beings like shiny, pretty things. In technology, nicely designed interfaces are perceived as easier to use than non-designed ones. This doesn't mean your project should be a beauty contest, however. For example, imagine if Google's screen design utilized rich imagery and elaborate screen transitions. While this might be appealing in another setting, it would be a complete nuisance on a search screen. For Google, and indeed many others, the most beautiful screen design is often the simplest.
It's worth it
We know very well the pressures on a new project to quickly get to work building something. It's unfortunate when steps like user research, prototyping, and user testing are the first things to go when budgets and timelines tighten. The irony is that these will often save time and money in the long run, and ultimately keep you from unwittingly rebuilding a merely better-looking version of what doesn't work.