I’m making the case that the most dangerous phrase in a design process is “it seemed to test okay”.
If it didn’t solve a problem than why it is there? If it doesn’t help people get something done – the thing the product is intended for – then it could just be getting in the way.
If you find yourself saying “it tested okay” you’re probably testing too much. You’re not really learning. It’s a signal that you might be testing the wrong things or asking the wrong questions. It could be that there’s too much noise in your design.
It’s better to test with less. It’s easier to discover what to add than what to remove. Design research suffers when we try to test too much. It’s easy to reach a feature or content plateau where things test okay but don’t help us get to insights, and ultimately, better solutions. This is the Inverted-U at work or an example of the “flat middle”.
Try removing anything you find yourself saying “tested okay” and test again without it. Do people miss it? Does the thing you’re testing work better without it?
Try to see things without introducing too much of your own context. People have enough context in their lives. We have to work hard to understand this complexity through the design process. Introducing too much design, too soon, makes this harder.
You only find out what’s really missing when it’s not there.