Usability Testing

Usability testing is the cornerstone of UX design

You cannot do user-centred design without engaging users. (It’s all in the name!) At Flow we run hundreds of usability tests every year, and we make sure that every member of staff is trained to run a successful test.

E-commerce websites that make changes based on usability testing typically see sales uplift of 83%. That’s well worth the money you spend on doing the testing.

“Watching your usability test of our site was a new and quite shocking experience. As we watched, we realised how we can boost conversion rates. We’re going to tackle the design process differently from now on!”
Andrew Smith, Yuppiechef

We can test all sorts of things:

• Live websites, or password-protected sites in a live environment.
• Clickable prototypes, which we can generate in Balsamiq, Omnigraffle or Axure.
• Mobile apps or mockups. We have a custom-built mobile testing deck, so we can record what the user is doing on the phone or tablet.

First, we work with you to write a recruit brief

It’s really important to recruit the right people. If you get the wrong kinds of users to come in, you’re just wasting everyone’s time. So we’re quite particular about this bit.

We will recruit the users for you, and make sure they show up at the right times.

The user interview: not a focus group

A Flow consultant sits next to the user. It’s a one-to-one interview. Sometimes the consultant takes notes while doing the testing, sometimes we have a note-taker in the observation room.

It’s all about observing what users do. For example, the user can’t find the CHECKOUT button. We let them search for it for a while, and observe where they are looking for it. We ask them questions like, “What’s going on now?” to get them talking about what they’re doing. If they really can’t find the button, we’ll eventually show them where it is.

Now the user will probably say, “Oh there it is! They should make it red so that it stands out more.” We are really not interested in the fact that they think it should be red. The user is not the designer. What we are interested in is the fact that they couldn’t find the button. Users can show us the problems, but are rarely good at coming up with solutions.

It happens at our lab in Claremont

We’ve worked in usability labs all over the world, and we built one that we all wanted to use. We also did some user interviews with our own customers, so the lab caters well for teams that want to observe usability testing. Here’s a map and directions to 18 on Cavendish.

We can also take our mobile lab out into the field. This means we can run usability tests with users in their own environments when it’s useful or relevant to the research.

We record everything using Silverback, while at the same time broadcasting the session to our boardroom.

Meanwhile, the clients are in the observation room…

Clients can watch what’s going on through a one-way mirror from the room next to the lab.

Or they can watch the broadcast from our boardroom, while talking and sketching and shouting: “I can’t BELIEVE he’s not seeing that button! It’s right there!”

Afterwards: a debrief session or workshop

We go through what we learned in the tests. For each issue, we create precise, actionable recommendations for the short and long term. For key areas we make sketches showing how the recommendations might look.

We can present results back in a comprehensive report, or a Powerpoint presentation. Or we can discuss the results and design ideas in a workshop with your team – for quick turn around and closer involvement.

Rapid Iterative Testing & Evaluation (RITE testing)

Doing your first usability test feels like a big step. But once you’ve got used to it, you can start doing rapid iterative testing, which gives great optimisation results for specific problems.

It works like this:

• We organise 3 rounds of usability testing, with 4-5 users per round.
• We leave 2-3 days’ gap in between the tests.
• During the gaps, we evaluate what the tests showed us, and revise the prototype.
• By the last day of a RITE test, the session tends to last 30 minutes rather than an hour because the users are just flying through the interface. When you give a user a task, they perform it without thinking, and look at you as if to say ‘duh, can’t believe you are asking me this’. That’s when you know your interface is ready to build.

We can test your own prototype, or design the prototype for you. The quality of the results are highly dependent on the quality of the prototype.

Ready to run a test? Get in touch!