Sometimes you find yourself writing a long email response to someone and think to yourself, “This should be a blog post.” Here are a few things I’ve learned doing usability tests on library websites, specifically Skokie Public Library’s website.
1. Test early and often
We actually started testing quite early using very barebones prototype versions of the site (one was a blank page with nothing but site navigation on it). Usability testing works best when it is fast and iterative. Do some tests, make adjustments, test again.
2. Do shorter tests, then longer tests
When we got closer to launch, we did at least four 1.5 hour sessions of quick (10 minute or so) usability tests out on the library floor (set up a table with a laptop, put out some candy or some other incentive and invite people to do the test). From these, we reached out to a smaller number of really engaged users to do longer (30 to 45 minute) tests. We scheduled a time and gave them a small incentive for helping (a Starbucks gift card or some such).
3. Have two people at least
You need at least two people running the tests. One person to be the MC or guide and the other to be a note taker. We also used screen recording software, but it ended up not being super helpful.
4. Scripts and questions
Be sure you have a good script prepared (we worked off of Steve Krug’s). And then define the tasks you want people to perform. One thing I’ve noticed that we librarians tend to get hung up on crafting the perfect task. I have to really encourage people that we just need good enough tasks. The actual benefit of the usability test is seeing real people who are not you use the thing you’re making. I have found that the specifics of the users completing a task or not is kind of secondary. That said, if there was a particularly contentious or hard to make decision along the way, that is a perfect thing to craft a test around (“What should we call databases?”, “Will people know what Downloadables means?”, and so on)
5. Normal people
My #1 rule when it came to thinking about website usability was that no one was allowed to claim to know what “normal people” would think or do until we actually sat down with normal(ish) people. Another fear or concern that often has to be addressed is working with a skewed sampling of people “because they’re already in the library” or “they already use the website”. You could try to get a pure sampling of your key demographics, but at the end of the day, the people coming into your library will surface issues you haven’t seen. Also, if you’re not designing the site for your library patrons, then who are you designing it for!
Oh, and you should probably have a screening question for people who approach you in the building to do the test, something simple like “Do you use our website regularly?” or “Do you use the Internet regularly?” will suffice. If someone isn’t comfortable using a computer, they won’t be much help in the test.