@brentlarue on Jan 27, 2015
Welcome back to the insider’s track of General Assembly’s Product Management Course with George Favvas. Last week we learned about features, user stories, wireframes and storyboards. This week is all about user testing, a key tenant of user-centric design. In this session you’ll learn when and what to user test, how to conduct a user test, and understand user testing best practices. Let’s get it!
First, why do we even bother with user testing? User test provide us with qualitative data that tells us how users perceive the product and value proposition, what is blocking users from accomplishing the set goals, what are opportunities to solve customer problems that were not thought of, and saves time and money by refining features before implementing them fully.
Second, when is the right time to user test? This answer may not be what you expect, but you don’t need a single line of code to user test. As soon as you have 2-3 use cases defined, you should begin user testing. In class, we sketched a fictitious app onto paper and broke into groups to test our paper prototype. In the spirit of the lean movement, you should create a product with the least amount of work which provides the greatest feedback. A paper MVP is a good place to start. Don’t believe me, try it yourself.
Depending on the maturity of your product you may want to test different things. Your usability testing should be in line with the same metrics you are trying to move along your funnel. If you are still working on the acquisition piece of your funnel, you may not want to user test the shopping cart experience. Be mindful of your funnel, product maturity, and customer adoption.
“Elaborate usability tests are a waste of resources. The best results come from testing no more than 5 users and running as many small tests as you can afford.” ~ Jakob Nielsen
Before you ever get into user testing, you must first recruit users. The method and approach to doing this comes in many flavors. You can expect to need at least 1-2 days before the test to recruit your users. Depending on your budget, you can use an agency, craigslist “gigs” section, or your own social network. Depending on your approach you will need some sort of screener to ask qualifying questions. To encourage participation, offer $50-100 Amazon gift cards for a 45 minutes session. These are super simple and cost-efficient to distribute to your participants.
Now for the test itself. Be sure when performing the test, be sure to take notes or record video. Some sample questions include: What do you think this product will do for you, would you be likely to continue on, why? While conducting the test, be sure not to include steps you feel are obvious. Pause on each wireframe or screen and solicit some responses. Point to specific aspects of the design, ask what they think it means, have they seen anything like this before, or what would they do here. You want to draw out gaps int he design or opportunities to improve. Consider asking if anything has surprised them, if they are seeing what they expected, what they expect to see next, and what they think the purpose of specific content is. If they get stuck don’t rush the test, allow some time for contemplation. Get comfortable with a little silence and then help them along. Long pauses could also indicate areas of confusion.
Lastly, how many users should I consult for testing? The answer – FIVE. Not surprisingly, zero users give you zero insights. With only a single user test, you can learn almost a third of what there is to know. The difference between zero and one user is huge, so there is really no excuse for skipping user testing. With the second user, you’ll uncover some overlap and some new insights, but not to the same degree you did with the first user. The third user will start to solidify feedback and provide relatively less new info. As you add more users, you will learn less and less. After the fifth user, you are wasting your time observing the same findings and not learning much new (consider The Pareto Principle). Check out the original research to learn about iterative design, why not test with a single user, and when to test more users.
That about wraps it up. Good luck with your users tests. Like and share below!