I love the idea of this, but time is such a weak factor. Would it not be more sensible to base it off total time used, or number of app launches?
I could install this and only use it once for the whole month.
An awesome first stab though!
It would be nice to combine both...
So if you use the tab bar (frequently), the icons would evolve, but if you were to leave the app for a certain length of time (a month for example), the icons would regress to their original state.
If you were to go off number of app launches, you may still hit the launch limit in the first couple of days, then not use the app for another month.
Really cool idea
Great thoughts, there will be a ton of variables and edge cases with this if it is intended to be truly "smart". I would absolutely love to see a complex GUI, think hundreds of controls, use this to help onboard new users and increase complexity over time, hours, launches or all of the above.
Yeah I think that would be a better approach. Also, may need some kind of reset, if the app hasn't been loaded in 6+ months, revert to state 2.
The rest is a nice idea. I can imagine the mechanics of how and when to reduce or increase complexity could be an entire project on it's own!
This feedback is spot on. Progressive reduction is one of my favorite things to have come out of LayerVault, but it should always be based on familiarity rather than anything else, and I think that comes with use rather than time.
Agreed, amount of use would be a much better scale to peg the changes to. Otherwise really well done, love seeing this stuff get codified. Nice work.
The cynic in me says this should be a user controlled setting. But then again, I'm not a fan of software presuming user habits..
Progressive reduction is one of these elegant ideas that really appeals to designers and product managers, but actually doesn't work that well once you actually try to apply it.
I am a fan of progressive reduction as an idea, but I think there are two keys for it to actually work.
First, the linked designs are excellent in that the dominant visual element on the tabs is unchanged - each iteration has the same circular icon in roughly the same size and place. This is key as users use shapes, size, and color to find targets in the corner of their eye, and you do not want your reduction to fight that.
Second I think this is only a good idea in apps people use frequently. If a product is used many times a week or day (such as email) the icons can be learned easily. But if a user only checks in twice a month (such as an analytics dashboard) then they may never become "fluent" with a text-free button.
Superb idea. This is exactly the kind of thing that cheers me up on a Friday.
Very cool concept, but still not sure about it. If a user only ever used one function a lot and the others are rarely used, once you hit the last level of reduction I think there would still be some issues with the lesser used functions.
I really think this concept is outstanding. Obviously balanced with testing, and seeing actual results. But none-the-less, I think it could be a great idea. Like others have mentioned, I'd love to hear a real life case study about this being used successfully.
This is great! My only concern is that fact the user will be confused when the shrink will happen, and they will try to find a way to get it back. From the other hand they just might think it was an update :)
Are there any case studies with this technique in place?
Loving the idea, I've also been investigating over this area of UI. My only concern could be that users might use certain sections/icons more often than others. In those cases, the unused icons become like some sort of ambient noise that the user is not perceiving, therefor, not learning their shapes and linked action/behavior.
Over usage, the progressive reduction mechanism could understand that the user is becoming more familiar with the whole app, while in reality certain icons would still be unidentifiable by the user without proper labeling.
I've been wondering how something like this could work by tilting the phone (using the accelerometer) instead of over time. The user is presented with the icons without labeling in normal position, and in case he's not familiar with any of the shapes, he can tilt his phone UP/DOWN to reveal the labeling. Not sure of potential downsides of this approach (discoverability?), but would love to hear your input.
This assumes that your users' use your app on a regular basis.