Testing and the Design Process

Now we are going to look at why testing is such a vital part of the design process.

Testing can help you check whether your design is working for people of all genders, if it presents any specific groups with problems, allowing you to gain feedback before launch. It allows you to run scenarios, testing any solutions you’ve put in place to tackle edge cases and gain the opportunity to discover new ones before going public.

The Importance of Testing

If we think about the Carem voice interface, testing would be a crucial part of the design process. The usability of the device is one of its key features, how easy is it for stakeholders to get the information they need when coming to the device with a problem? Much of the success of a voice interface lies in how well its conversations have been designed and the most effective conversation design for Carem could only be achieved through testing the device with stakeholders. There has been much criticism around the ways in which voice interfaces inadequately respond to abuse through poor conversation design (1), which in turn perpetuates gender inequalities, these are the kinds of issues that can be mitigated against with robust testing.

Testing allows for you to both intervene when features don’t work and celebrate when they do. Here are some suggestions for how you can consider gender and embed inclusivity into your testing:

Test Throughout the Process

Implement testing as early as possible, not just in the prototyping stage. If you’ve engaged in inclusive research and participatory design, you may have already gathered some useful preliminary test data or insights about how your product might work differently for people depending on their gender. Testing early on will help you foresee problems that could arise at a stage where they are harder to fix.

Test With a Range of People

It is very important that the people you are testing your products with represent a range of backgrounds, gender identities and abilities so that you can best understand how your product serves different groups. Even if you have designed your product for a specific social group, test it with other groups of people, especially with individuals that might experience specific vulnerabilities when engaging with your technology.

Consider Accessibility

Prioritise accessibility when testing your technologies. For example, if you’re creating a web-based platform, have you considered how someone who is visually impaired would engage with it? If you’re building a product that requires physical input, have you considered how someone with mobility issues engages with it? For designs to be inclusive they should be functional for all abilities. To learn more about accessibility you may want to check out the CCI’s course [Creating Accessible Interfaces](

Budget for Testing

If you have the means to, it’s really important to pay people for their time when testing your technologies. The feedback you receive will be vital to the project and teams should be acknowledging that by compensating people for their time. When people are paid they feel more valued, which will only lead to more thorough feedback. This point is important to consider in light of the gender pay gap and the way that women, trans and non-binary people are continually undervalued in the workplace.

Design the Testing Experience

How the testing experience runs is up to you, and there are many ways to go about it. You can run moderated or unmoderated sessions. Remote or in person. Design the testing experience you need, considering: Where do you need feedback? Are there any specific issues you’re trying to overcome? Make sure that the experience you design is a safe one. No one should come to any harm from situations you put them in. If you are running sessions with guided questions, design these sessions efficiently, ask pointed questions which will help you to categorise your findings afterwards. Equally, you might choose to let people engage autonomously with your product.

Keep a Testing Log

In the same way, you might think about creating a log for feedback, make sure that you keep track of the outcome of any testing sessions. Think about how you can gather data into material during sessions.

What are your thoughts on these suggestions?

References:

1. Nick Statt, 2019. AI voice assistants reinforce harmful gender stereotypes, new UN report says, The Verge.