Forum Moderators: open
http://www.pixelsurgeon.com/pages/interview/design/jakobnielsen/index.html [pixelsurgeon.com]
I've found that even un-savvy users can find links no matter what color, with or without underlines, if they're in a clearly defined navigaton bar in an area of the page where they expect them to be.
I just read a study where they asked people to try to navigate a site where the links were "disguised". Basically, most people tried the text in the left hand margin. This goes to show, as you point out, that most users will find links in a left navigation column, regardless of how the links are presented. On the other hand, they will do a lot worse if you put the links in the righthand margin.
Logically, links should go in the right hand margin because, on a PC and Mac, that's where people have to go to scroll and it would reduce mileage. So you can
1. have standard links on the left
2. non-standard links on the left
3. standard links on the right
4. non-standard links on the right
Numbers two and three will work because they "skid" the "standard" but retain enough of it for users somewhat familiar with the web to extrapolate and use the site. They also can present significant advantages in terms of visual presentation or usability.
However, number 4 will present more problems. Again, if it looks like an interface, users will figure it out, but they need enough clues, enough familiar characteristics to make it work.
The main thing is that it is entirely testable. Put the navigation on one side then on the other. Ask ten users to accomplish a task with each layout. If there are no differences in the time it takes or the error rate, your design is a success. If there are differences, the design needs tweaking.
Tom
I can't see how the study you referenced can in any way correlate to the real world.
Glad you asked, since I completely misrepresented the methods in the study, but not the conclusions.
Actually, what they did was give users a grid and then some cards representing common objects such as shopping cart, merchandise links, site search and have them place them on the grid to see where people expect such items to appear.
I think the point I was trying to make is still valid. If things are *where* they expect them to be, they can vary more from *how* people expect them to look. If you vary both the expected look and expected location, people will be completely lost.
You may disagree with the study methods, but please don't judge it based on my inaccurate capsule summary. Check it out at
You may find these interesting too
If there are no differences in the time it takes or the error rate, your design is a success.
There is a third factor that's good to include in testing - the subjective impressions of the test subjects.
For example, I've seen font readability tests where one font tested better in both reading speed and comprehension -- and yet the average subjective "take" was that another font felt easier to read.
If people don;t like what you do, even if it tests well, it's a problem.
third factor that's good to include in testing - the subjective impressions of the test subjects.
Good point. In the second article I mention (actually, I should check on that) they found no significant difference in time/error, but big differences in subjective impressions.
Tom
