Bite-Sized UX Research

What About Card Sorts?

Conducting a card-sorting exercise early in a project is desirable when the concept space is not well understood. Early on, you’d typically perform an open card sort, which may require more time and resources—for recruitment, implementation, and analysis—than you have.

Although it is possible to design, recruit, conduct, and analyze a card sort in a matter of days, the question remains whether you might spend that time on another task, offering higher impact to the project. However, there are times when nailing the information architecture is the most critical element for the success of a project, and a card sort would be the best use of your time.

What to Deliver?

Although your time may be short, don’t abdicate responsibility for an information architecture to others. Do wireframes, even if you show only some elements in detail. The critical features of your project will drive your choice of what elements to detail. You might spend your time designing the elements that are most important to the success of a project, leave less important elements to the development team, and iterate the design of those less important elements later on, when you have more time. Alternatively, design and test complex elements that it would be costly to redesign and reimplement at later stages in a project.

As Time Goes By

As a project moves out of its early stages and into the implementation cycle, you can start shifting your attention from requirements to refinement. At this stage of a project, useful activities include

• user walkthroughs
• closed card sorts
• usability testing
• definition of metrics

User Walkthroughs

You can do small-scale user walkthroughs of wireframes, a paper prototype, or a stack of screen shots for very little cost. The aim is to generate feedback from real users—or, at least, realistic users—on the design of features at a relatively early stage in a project—that is, before too much development work has taken place.

This activity doesn’t necessarily have to be a big deal. Take a stack of paper printouts to a local cafe and ask people if they’ll walk through a task or two with you, in exchange for a coffee or snack. Don’t take up too much of any one person’s time. It’s better to get several different people to try out a feature or two rather than possibly annoying people by demanding too much of them. They’re trying to relax after all!

Iterate this process if you can.

Closed Card Sorts

A closed card sort is a little easier to conduct than an open card sort— and, in my opinion, a lot easier to analyze. Using decks of cards representing a site’s structure as you’ve designed it, ask people to find their way to particular low-level pages. You can plan your card sort based on the major user tasks, critical features, or areas of known ambiguity.

You can easily conduct a closed cart sort just about anywhere, and it should take only a few minutes per task. For each task you test, note the path each user takes—that is, the card a user selects at each step—and whether a user successfully located the low-level page.

You can carry out a test like this in a single day, including its setup. Since time is short, focus on task failures and look for common mistakes participants make. Report only such issues, but make the full test results available to your project team. You don’t want to be the bearer of only bad news. Instead say: “A lot of stuff worked well, but here are some things we need to look at changing.”

Usability Testing

Using whatever version of a Web site or application you have available—whether wireframe printouts or a low-fidelity prototype—conduct a more formal usability test, during which you ask each participant to complete all tasks. Although this can be time consuming, a usability test with six participants takes just a few days and can provide valuable insights for your project team.

You can take this opportunity to test your whole design solution. Alternatively, you can concentrate on high-impact features, according to your time, people, and budget constraints.

As a finished Web site or application becomes more tangible, continue doing iterative, small-scale usability testing. There will come a time when you can’t incorporate the findings from your testing into the upcoming release, but your recommendations can form the basis for a subsequent version.

Define Metrics

If you’ve joined a project very late, there may be very little you can do to influence the finished product. However, you can still add value to the project by defining a set of analytics that can help the project team make design decisions during the product lifecycle. So, get the team to record page-completion times for a multi-step transaction process, make sure you get good search logs and reports, and give yourself an opportunity to learn something about how customers are interacting with your site.

Don’t be disheartened if your project team can’t incorporate your changes straight away. Start laying the groundwork for the next iteration of the product or service and stay ahead of the curve if you can.

Summary

If you don’t have the resources of a large UX team, with the budgets and timelines to undertake the ideal user-centered design (UCD) or activity-centered design process, you can still make a valuable contribution to a project. Undertake small, tactical, iterative user research activities throughout the course of the project. Focus your efforts on the areas of greatest impact, and produce documentation that your project team can understand and use efficiently.

If you demonstrate value through a series of small, high-impact UX activities, the extra budget, people, and timeline flexibility you need will eventually come your way. Then, you can come closer to implementing your ideal UX process. 

I’d like to thank Ruth Ellison, of Stamford Interactive, Daniel Szuc, of Apogee, and Russ Unger, of User Glue, for motivating me to write this column and for their input into the ideas it presents.