Exceptional experiences are the expectation
Let’s face it, we are all so spoiled by technology. It’s fast, it’s efficient and it’s right at our fingertips. We call an Uber with a tap when we need to get somewhere. We buy a bunch of clothes from Asos quickly and with fair confidence, it’ll fit us. We open Google Maps to make sure we get to our destination on time.
Today, exceptional experiences are the expectation. User experience will soon replace price, products and services as the key brand differentiator. Leading companies are putting all their eggs into the basket of managing the end-to-end customer experience and delivering that exceptional user experience for their products.
Now what I talk about here could be applied to the user experience of virtually any product, including physical products like hardware tools, building a piece of Ikea furniture or buying a ticket for a train. I’ll mainly be talking from the aspect of software like websites and apps here but, keep in mind you can apply these principles to improve any product or service.
What is usability testing?
One of the main ways we improve end-to-end customer experience is Usability Testing (or user testing). It’s the process where we evaluate the design decisions we made for a product against a targeted set of users to test if our assumptions are correct. We should always test our assumptions.
Why do you need to do usability tests?
Testing will help us identify problems people have with specific UI/Interactions and reveal issues to completing tasks. It involves quite a bit of preparation and analysis, but it’s one of the most valuable research techniques. We cannot assume to know how everyone will use and feel about using our product. Put it to the test!
When should you do testing?
Usability Testing can be done at many points when creating or improving a product, some suggestions could be:
- At the wire-framing stage where we have a clickable prototype that a user can interact with.
- At the design stage where users can engage with a clickable prototype.
- After the product is developed.
How many participants do you need in a usability study?
According to research, testing with 5 people lets you find almost as many usability problems as you’d find using many more test participants. The vast majority of your user research should be qualitative which is geared at collecting insights to drive your design, not numbers. Keeping the number of testers below 10 also helps increase your return on investment and save cost. Testing costs increase rapidly with each additional participant, and there is the point of diminishing returns. I personally like to do more, between 5-10 if the budget allows for it, it’s good to allow for any chance of non-ideal or uncooperative participants.
How do you do usability testing?
I find many tend to overthink and over proceduralise this. Keep it simple. Here is a couple things you should do:
- Test one participant at a time!
- Clearly set the goals/objectives of the testing.
- List the tasks/functions/features you are testing – ie. signing up with an account, booking a ticket etc.
- Always record your session so you can refer back to it.
- Encourage the participant to think out loud when using the product.
- Don’t lead and instruct the participant on how to do things.
- If you can’t have more than one facilitator, focus on facilitating the session rather than logging feedback. We can always do that later since we recorded the session.
Quantifying the testing data
You could have potentially hundreds of issues arise from a usability study, that’s a lot of data to process. How do we know which issues matter, and which we should address first?
I will defer to Mr Carlos Rosemberg’s article on ‘Turning Usability Testing Data into Action without Going Insane‘, it’s the most comprehensive guide I’ve found that teaches how to turn the issues at hand into prioritised solutions you can action.
Keep on testing!
We’re always too close to our own products and that makes it easy to make misguided assumptions. Competition is high, and customer expectations are too. All you need are 5 participants for a consistent testing environment. Why not try testing your product? And then test again, and again…
