A hypothetical discussion between two designers reviewing their design work. And some hypothetical effects for some hypothetical metric (ex: leads). Here we go:
- Hey, want to review my UI screen and make it better?
- Sure thing Bob. As a start why don’t you move that logo more to the right? 0% Effect
- Awesome feedback. Think I should make that header photo bigger down as well?
- Yeah, huge photos are nice. Actually make that background animate. -11% Effect Oh, and see that tiny signup link over here? Please also make it more visible, by turning it into a button +7% Effect since it’s a primary call to action. I’ve seen some really nice ghost button implementations out there -2% Effect – perhaps you could consider that as well?
- Thanks, anything else?
- Your page is missing some social proof +3% Effect, trust seals 0% Effect and a how-it-works section -9% Effect. That’s all I have for you.
- Thank you Joe.
All design changes are risky. They may have negative, positive or no effect at all on all sorts of key metrics. If we were to add up all those hypothetical changes above, we might observe a -13% net effect (along with time and money being spent on the work itself). And yet, such design discussions happen regularly and lead to all sorts of UI decisions being made.
One simple way to make sure that only positive changes are rolled out onto production is to measure every change – ideally as a controlled experiment.
Another powerful way of increasing our chances of making better design decisions is to use past test data when designing anything. Past test results carry probabilities and therefore lower the risk of suggesting changes with negative effects.
Share Your Thoughts
How do you increase your odds of making better design decisions?