Low-Cost Usability Testing: Why and How
This is a guest blog post by Laura Orban.
Usability testing should be part of what every website owner does on a regular basis. In my experience, few actually do it. Or if we do it, it’s only done a week before a major redesign launches.
Part of why we don’t usability test is because of what we think it has to be. We imagine an all-out usability lab complete with a one-way mirror with 10 people sitting behind it. We imagine thousands, or even tens of thousands of dollars being spent.
If you’re responsible for the software that lands airplanes or that heart surgeons use, by all means, do that. If you are one of the millions of websites on which people’s lives don’t depend, you can try the low-cost way.
We should all be testing. Businesses large and small, non-profits, even municipalities and public schools should test.
Why Do It
We can (and should) test our websites early and often. One of the most common objections to low-cost usability testing is that it’s not going to yield quality results. That’s not true. Subscribers to that belief don’t tend to do the expensive kind, they tend to do nothing.
Getting some feedback is always better than getting none.
Usability testing expert Jakob Nielsen tells us that low cost testing done iteratively is better than doing nothing and can actually yield very good results.
Testing with just 5 users will find 85% of the problems. While you may need 15 to find all the problems, it is better to test with 5 on 3 different occasions than it is to do one test with all 15 and stop.
How To Do A Low-Cost Usability Test
We’ll go through each of these steps, one at a time:
- gather your stakeholders
- document the goals of testing
- create test questions/scenarios
- find your participants
- conduct the test
- review results and decide what to fix
1. Gather your stakeholders
It’s important to get the people who are invested in the project involved early. You don’t need every person, but you should have a representative from the critical areas like: leadership, user experience, operations, development, content authoring, metrics and your business owners.
You don’t want to be surprised right before the test because someone who needed a say in the plan didn’t know it was happening.
2. Document the goals of testing
You need to document your goals. What do you hope to learn? Is there a new feature you want to test? Are you comparing 2 designs? Do you want to know why users aren’t completing a critical task?
If your site is live, you should look at what your metrics are telling you at this step. Web analytics can tell you what users are doing, usability testing can tell you why. Use your metrics as input to your test goals.
3. Create the test questions
The goals of testing inform the test questions and scenarios. They are not the same.
The test questions are the things your team wants to have answered. But you aren’t going to sit someone down in front of a computer and ask them to tell you outright if, say, your categories make sense. You are going to give them a task which will allow you to see if the categories are making sense.
Your test questions and scenarios will consist of the things you want the user to do (maybe complete your sites most important tasks) the results of which will yield answers to your test questions.
4. Find your participants
This is where the expense may come in. You will probably have to compensate participants for their time. It shows that you respect and appreciate their input.
This is compensation for their time and not compensation for specific feedback.
Renowned author and low cost usability testing expert Steve Krug tells us that when it comes to finding participants, “recruit loosely and grade on a curve.”
That means that for the sake of speed and money, you’ll do just fine by selecting 3-5 participants who are as close to your typical user as possible, even if they are not an exact match. Just keep that in mind when reviewing the results.
5. Conduct the test
Your set-up is relatively simple. You need a computer, 2 chairs and either a video camera or software that will capture your screen. It will be just you and the participant in the room. The video or screen capture (which must record your voices) will be used to show others the results. You don’t have to take as many notes when you have it recorded.
Steve Krug offers an excellent script you can follow. You want to make sure you hit the following points:
- let the participant know they’ll be recorded and get a release signed
- tell the participant that you are testing the website, not them
- let the participant know that you want honest feedback
- make sure the participant knows that s/he can stop the test at any time for any reason
It’s important to make the participant feel comfortable and to respect their rights.
Start off with a few casual questions: what they do for a living, how much time they spend online, favorite websites, etc. This will help him/her to relax, and also possibly give you some insight into what they’re like.
Tell them their task. Make sure their starting page is ready to open. Ask them to think out loud. This is critical. Your job is to learn why they are doing what they’re doing. You may have to prompt them (“why did you click on that link?” or “what were you expecting to find here?”).
6. Review the results
This is the last step. You and your stakeholders will review the results together. You’ll decide what you’ll try to fix based on what you saw. There is no 20 page report. There is a 1 to 2 page write-up documenting what you saw and what you’re going to do about it.
Steve Krug recommends picking one morning a month for user testing. Test 3 participants in a row, and have lunch as a team to review and decide what to fix. It’s that simple.
What Could Be
We’ve all had the experience of using a website that is so bad, we can’t imagine the people who own it have ever tried to use it. But it’s easy to use our own stuff. We need an outside person to show us why the terminology we know doesn’t make sense to the average user.
Testing early is important. Draw your design on a piece of paper and test that before you spend hours coding.
What did I forget? What questions can I answer?
Laura Orban is an internet marketer, movie quoter, mom, lover of big rescue dogs, wife, compulsive volunteer, problem solver, and happy. Visit her blog at LauraOrban.com and tweet her at @LauraOrban.
You must log in to post a comment.