Sometimes when you design things you become blind to the fact that your graphical design does not actually work all that well. It can happen that you fail to account for some corner case or you can fail to account for missing contextual information etc.
How do you ensure you are in fact designing it right and the design works as intended? When doing some user testing how do you ensure you do not taint the test with your context that then does not always exist? Or we might ask teh wrong audience. What other options do we have?
Some examples:
Image 1: Sometimes you just don’t get the design as presented.
Both of the designs above probably were perfectly valid in environment where they were designed, just not when they were applied. It is just that the selection highlight probably worked quite well when the designed menu had more than 2 items. Well tough luck in the version i was presented they weren’t present. Likewise the symbol if perfectly understandable if you happen to see its counterpart, but alone it does not work as well. Unfortunately in practice this comes up every day since the icon are not all next to each other.
Answer
Testing with real users. As simple as that, there’s no other VALID way.
You can test with fellow designers, but it would be completely biased and quite probably you’ll get opposite results to what “real life” would bring, so avoid that, unless this is a “design for designers” (such as a competition, or skills display). I’m not a graphic or web designer per se, I’m an UX Developer, and I’d say 90%+ of Behance or Dribble stuff won’t work in real life, no matter how beautiful it looks.
The thing is you need to do a trade-off between visuals and UX. Basic UX rules tells your examples are not a good fit (as you have found out already). However, you also say that it works in context (does it?), and UX is all about context, so you might be right, it works in context.
About your specific samples: if by highlight you mean a hover state, then the interaction will transmit an affordance, so you won’t have trouble with that. However, if this is a static state (such as an active page), you’ll probably need to add something else to transmit where the user is at, and why that link has a different color. In UX, the recommendation is to underline links and be a happy camper, but well, here’s where the above mentioned trade-off comes to play.
With regards to icon, there are very few universal icons (that is, icons which everybody understand): print, search, home and few basic icons like that. Everything else is subjective. If I take a look to your icon, I honestly have no idea what it means, and I really tried hard to figure it out, with no avail. Something related to a person, because of the “head”, but really no idea. In UX, you add labels to all icons, it’s a basic principle. If in your context your icon has labels, then great. Otherwise, you’ll be in trouble.
In short
There are many many rules to follow if you want to validate a design with users. But even if you follow them, you’ll always need to test, that’s the only real validation.
Additional Reading:
Attribution
Source : Link , Question Author : joojaa , Answer Author : Devin