We got together with UX designer Mariano Andres Garcia and UX professor Guiseppe Getto to talk about how you can optimize your usability research to get the most insight out of it and get beyond the obvious issues to see the more subtle usability problems. Below is a recap of the Q&A from the webinar.
Q: A lot of times I hear people say, ‘I don’t have time for usability testing,’ or ‘I don’t have time to view the results.’ What would you say is the right amount of tests to run? What’s the minimum number of tests you can run to get actionable feedback?
Mariano: First of all, in terms of not having enough time, nowadays its super easy – especially with tools like TryMyUI, you can set up a test in literally 5 or 10 minutes, and you can even have somebody watch the video for you and give you the actionables, so there’s really no excuse not to test nowadays.
Really there was never a good reason not to test because at the very least you can ask the person sitting next to you, or your wife, or whoever you have nearby to tell you what they think and try to see if they get what the product is for.
As for the number of users to test with, usually the first person that you test with will give the most insight. They will find the most critical, most important errors in your platform and then the second one will have some of those errors again and give you a bunch of other ones, and then as you keep on adding people to the equation, you will continue to confirm what the first couple of testers gave you and maybe find out another thing here and there. So as long as you’re testing at all, you’ll definitely be getting something useful out of it.
Q: There’s a lot of talk about ‘mobile-first’ design. Do personas change at all when you’re coming from a mobile-first perspective or when you’re trying to optimize your testing for mobile-first, how do you change your personas for that?
Guiseppe: The personas are always tailored towards the application, and this is why I would say personas become more important in a mobile-first environment, because we don’t know where people are accessing our applications from. They could be on a tablet, they could be on a phone, if they’re in a state like North Carolina where I live, they might be on their phone in the car. A persona really gives you all that messy context. So I think they’re even more important in a mobile-first world because you need to know who you’re responding to. What is your design responding to, what is the user context; and again to my mind the best way to depict that is through a persona.
Q: Do you find that as you collect more data, personas evolve? How often should you update personas? Say I have a ‘Robbie’ and a ‘Susan’, but as my data is coming in it doesn’t all fall into the Robbie bucket or the Susan bucket. Do I create a new persona or do I just change personas to adapt?
Guiseppe: They should be updated pretty much continuously. So for this project I’m working on right now, those are really preliminary personas, we’ve done a first round of testing and we think we know that these are good personas for the application but we’re not sure. So they definitely evolve over time. And that’s exactly right, they sort of emerge from the data.
So for example if over the next year or so as I’m working on this application, if I test with 10 more people, and they’re totally different from ‘Robbie’, then that’s going to become a new persona, that’s exactly right. You have to weigh the different trends in the data, and you have to say, what is the statistically relevant trend? Am I going to say if 20% of my users are the more adventurous type, that I’m going to include that? And you can think of it in that sort of fine-grade detail. But the idea is that every trend that is important has a persona. And you can build a lot of them. I’ve heard of people working with IBM that have 33 personas, and that’s all of their products.
The other thing about personas I should mention is, if you can’t get out in the field to interview or to test right away, you can use your personas – personas do give you a lot of traction and enable you to say, ‘Well, we know that this user base is something that we’re targeting’ and you can use that to design, and give it to developers and say ‘Ok how would this type of persona go through the application?’ so it actually would save you time in the long run. But they do need to be updated, definitely.
Ritvij: I’d like to add on to this – the whole idea of using personas is really valuable for running usability tests and analyzing usability test results because now you’re no longer dealing with 100 individuals or 50 individuals. As you go on bucketing people in personas it’s essentially like running a kind of extrapolation on a set of data points and it’s really helpful because it gives you a good read on the situation and it also streamlines your analysis process. So this is a way to run usability tests and have something more concrete arise or emerge from them.