The only thing constant is change.  In User Experience, change is always good when it derives from research. Iterative tests have a lot to do with both change and research.

This is how iterative testing works. When a doctor tries to treat an illness, the first thing he does is searching for a cause of the symptoms.  Most of the times, he cannot be sure what the problem is caused by but he hypothesizes and starts a treatment. Trying any treatment is better than no treatment at all.  When the treatment doesn’t work, he tries to understand why, and tests another hypothesis of what can cure the illness. The whole process doesn’t stop until the illness is completely cured.  I know that the reality isn’t always as promising as the metaphorical doctor, but iterative tests are nothing more than setting hypotheses of what can cure the problem, watching the reaction to the applied treatment and improving it repeatedly until the problem is gone. Instead of people, we cure websites; instead of medicine, we use design.

How did we do it?

Iterative tests can be quite expensive because nobody knows when they would give any positive feedback. Actually, I bet you can do them infinitely.  This is how we did it…

The Client was a big e-store selling home furnishing, decorative articles and textiles. Its best-selling products are roman blinds. They are sewn individually for every customer using a fabric of his (or rather her) choice. There are thousands of different combinations of fabrics, finishings, sizes and mechanisms that a customer has to configure through the e-store.  We knew that a product page would be a challenge, and so it was. We decided to conduct iterative research with customers on the most important functionalities, like menu, product page, listing and fabric samples ordering. We had created low-fidelity mockups with core interactions only and went to the Client’s brick-and-mortar store. We spent there 5 days engaging customers in 20-minute tests in exchange for a small discount.

Every day was a different test and improve cycle. Each morning, we made changes to the mockup based on conclusions from the day before to start testing it in the afternoon. We tested until we had found some tendencies, usually between 6 to 7 people a day, which gave us 36 respondents  giving feedback on the same functionalities.

This was enough feedback to give us a memorable lesson in humility.

We tested the first 3 versions of some filters. It wasn’t too bad, but it wasn’t good either. We introduced some changes on the 4th day thinking that there was no way users wouldn’t have been able to use it the way it had been designed for… and the feedback on it was even worse.  Fortunately, the next change which was the big comeback to the simplicity and straightforwardness, caused positive feedback.

Here are some conclusions from our experience:

  1. People are willing to give you feedback, you just have to ask for it.
  2. No matter how much of an expert designer you are, the chances you would be right on your first try are rather small.
  3. Keep your mockups simple and with no visual design. Pictures and colors can catch too much of a respondent’s attention.
  4. Try to recruit people who have no difficulty in abstract thinking. There will be a lot to imagine for them.
  5. Keep your mockup in a sketchy style so that you won’t fall into the temptation of finishing every single detail of it. It doesn’t matter what color a button is, your aim is to make people see that it is different from a link. That’s all.

Iteration is the perfect way of getting to the right choices of design.  All in all, we don’t want something that only looks good, but something that works well for us. A good way to find it out is to test and iterate, even if you, as a designer, are about to find out that you are not as predictive as you thought you were.

Iterative tests are a big part of Lean User Experience approach. You can learn more about what Lean UX is from our post “What is the difference between User-Centered Design and Lean UX?”

Learn more:

Anna Lankauf

Interaction Designer at Divante eCommerce Software House

Share your comment