I am looking for some studies that examine how users perceive controls (e.g. navigational icons) that are rearranged based on most recently used. Also it would be interesting to see how rearranging effects user performance. My first thought that rearranging icons could be annoying because users may not know why the icons are moving and cannot remember locations of the icons because they could be different. On the other hand, moving the most recently icons to a more prominent places could make them more easily accessible especially when time matters (e.g. two most frequently icons appear on the first page vs. the second page). Currently I am trying to see if rearranging icons based on most recently used is a good or bad idea for the users.
It is well known that when people select icons or buttons, they aren't consciously evaluating that icon and then choosing it. They are usually going by memory.
The factor that most effects our memory of items to select is their spatial position. So if you move icons as people use them, you will end up subtly frustrating people. As an example, Microsoft Office did something similar years ago, where they would move menu options that you didn't frequently use. This caused a lot of confusion and they eventually abandoned it.
I would suggest looking globally at the most used icons, and then putting them in an order that makes the most sense. But keep the order fixed.
Use icons for:
- Commonly know operations (print, save, add, delete, etc.)
- Branding (let people know your brand and partner brands)
- For bullet-points, errors, questions, warnings
But remember, having a fixed position of icons/menu items is important.
In the second experiment, we kept the original pictures, but shuffled their locations on the toolbar. To our surprise, users really struggled with this. It really slowed them down, and, in several cases, they could not complete common tasks. (The icons were all visible, they just had trouble finding them in their new locations.) UIE 2006
Think about it -> at first it might be useful to read the signs of a shopping mall to find the electronics section, but once we know where it is we go there without first reading signs. http://en.wikipedia.org/wiki/Spatial_contextual_awareness
I agree with JohnGB and have also read a while back that users tend to remember the position of icons more so than the icon itself.
There is on instance where I can see rearranging icons by most recent to be of value. If you had a very long list, for example 50 icons, which required you to scroll down the page to see all of them, placing the most recent icons first would be more efficient than forcing the user to scroll down the page.
This would also make sense as once you get past a certain number of icons, the list becomes no longer scannable, and hence having recent icons in the first few positions would make finding the approriate icon quicker for the user.
I think this is an area where metaphor still wins over idiom. The metaphor is as follows: if I am building a bookcase or fixing a computer and I have several tools out and around me to help me complete the job, when I put one tool down and pick up another that first tool stays where it is and when I need it again it is where I expect it to be.
I think also there is another UX principle at work here that the system should not make a decision or assumption and move / close things that a user has placed and / or opened themselves. IPhone Safari does this by closing the least used tab when more than 8 are opened. Some tabbed interfaces make this mistake too by having two rows that switch position, or grow from one to two rows when enough are opened. PHP storm makes this mistake.
Allowing users to arrange icons themselves then leaving that arrangement fixed is almost certainly a much better solution.