Introduction

Web and interface designers are in a unique and powerful position to serve as user advocates. By incorporating usability and conversion principles into design from the beginning of the process, designers can help ensure the user-friendliness and effectiveness of the end product.

It sounds great, right? But how (you may ask) can designers do this? Classes and reading are one way, of course – having a solid understanding of usability best practices and research will naturally influence your design decisions. But getting from the theoretical to the applied can be a challenge, and best practices are not, in my experience, always sufficient. Ideally, each unique design is tested with representative users to validate whether it really accomplishes what’s expected.

Which is where this series of articles comes in. I’ll cover a number of quick, easy usability tests that designers can do in very little time and at low or no cost. You’ll end up with validated designs, better end results and a growing understanding of the complex, powerful relationship between usability and design.

We’ll start with a very simple test, and build from there.

Test #1 – Button Labels

Test category: Expectancy

What we’re testing: Button labels

What we want to know: Which version of a button label best communicates to users what happens after they click the button?

Why this matters: Button labels are a crucial point of communication. If users quickly and easily “get” what will happen when they click a button, they’re more likely to do so. If, on the other hand, a button label is unclear, confusing, or makes users feel uncomfortable, they are likely to hesitate or avoid the button altogether.

When to do this test: Conduct the test whenever button labels are being finalized or updated. This can happen at just about any stage of a design or redesign process.

What the test tells you: This test will tell you which one of two or more button labels does its job the best. You want a button label to support the audience’s mental model—that is, what they expect to see or have happen when they click the button.

What the test DOESN’T tell you: This test won’t tell you how the button will perform within the context of an actual web page or application. Users may understand a button label when viewing it in isolation, but when many other factors come into play the results may vary. For instance, users may not see the button or may not be sufficiently motivated to click it, depending on its placement, the other messaging on the page, what the user has experienced prior to the page, etc.

What you need: Several buttons identical in design but with different labels. These can be low-fidelity sketches or high-fidelity renderings, as long as the button design itself is the same for all the buttons.

How to set up: Arrange the buttons so you can show them to test participants one at a time—one per page or per screen—to minimize any cross-influence. Make sure you’re able to vary the order in which you show the buttons.

Who to get as test participants: When conducting any usability test, you should seek out participants who are similar to the target audience. This can be a small hassle, but it’s very worthwhile because you’ll receive more authentic, useful feedback.

How many test participants to include: For this type of test, 5-7 participants should suffice, assuming they are representative of the target audience. There are two benefits to including more participants, if that’s an option: you’ll be more likely to uncover subtler problems, and your results will be statistically more reliable. But for a quick validation check, 5-7 is enough.

How to prepare the test participants: You want your test participants to feel comfortable giving their genuine feedback. Here are a few quick things to tell them before you get started:
- Thanks for participating!
- You’re helping us improve a design, and we really appreciate your help.
- We’re testing the design, NOT YOU. There is no “right” answer to this test.
- We’re looking for your honest, gut-reaction feedback.
- Nobody will be offended or hurt by what you say. Everyone involved wants this design to be the best it can be.

How to prepare yourself: Have your buttons ready to show, and prepare a place for the test participant to sit. Have a notebook and a pen ready to take notes. Label each page of notes with the corresponding participant’s name, so you know who said what.

How to run the test: This is probably the easiest part! Here are the steps:
- Bring your test participants in one at a time.
- Vary the order in which you show the buttons. For example, show the first participant Button A, then Button B, then C. For the 2nd participant, start with Button C, then A, then B. In other words, mix up the order each time.
- Show the first button to the participant and say, “Take a look at this button. What do you think would happen next if you clicked it?”
- Write down the participant’s response.

If the response is very short or unclear, ask non-leading follow up questions such as, “That’s interesting, could you elaborate?” or “When you say ‘A big mess’ – could you describe that to me?”

- If the participant asks you a question, such as, “Is this button on a form?” do not answer directly. Instead, ask a non-leading follow up question such as, “If it were, what do you think would come next?” You end up sounding a little like a psychiatrist, but it’s important to stay neutral.

Make note of the questions. These can provide valuable clues with respect to how the participants reacted to the button, where they would expect to see it, and so forth.

- Repeat for each button.

What to do with the results: Compile the results and look for patterns. Ideally, one of your button labels will evoke the desired response from a clear majority of participants. If so, you’ve found your winning button label! If not, the labels may have been too similar to one another. Or, you may be farther from a solution than you first thought—which simply means another quick test, with updated button labels, is in order.

Is there any particular design challenge you’d like me to address in this series? If so, et me know in the comments.

This entry was posted on Monday, February 4th, 2008 at 2:04 pm and is filed under Usability, Web Design. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed.