AP UX Week '07: Creating the Adaptive Interface
[Live blogged notes from Adaptive Path UX Week 2007. Creating the Adaptive Interface" by Stephen P. Anderson]
At the base level we need things to function, we are focussed on tasks. There is a zone of "safe mediocrity." Above that we get to be more personalized and this is where opportunities for competitive differentiation exist.
We are not talking about radical changes from day to day. We are talking about micro-changes. We are not talking about MS "Smart Menu." Moving things around doesn't work. We are not talking about adaptive content like at Amazon.
We are talking about the interface itself. Simple changes that we can make right now to our applications. Think of this like a choose your own adventure novel. Think about components as dominoes. They can match up in different patterns and this opens up possibilities for different experiences.
Basecamp, for example, could track how many files I post with each message and expose the average number as the default for me. There isn't one right way, there are right ways for different people in different situations. Look for the right way for your audience.
The Dream Project
It was a group information management system. The user base was four people and they knew exactly who would use this. There were no constraints, they could buy giant screens, etc. "The application needs to work like I think" was the directive of the manager.
But...
Between four people there were:
- different computer literacy
- information needs varied
- display of information varied
- screen size varied
Solutions:
- custom layouts for context
- custom layouts for screen size
- different calendar views per person (month view v. two week view etc.)
Working on this he learned that it isn't that hard to build these custom interfaces.
If this happens we know that if/then else or What if, why not, if then.
2×2:
- Adaptive - Adaptable
- Content - UI
Examples, etc.
This is like a buffet, take what you like.
Adaptive text input boxes that expand to fit. This let's you avoid having to decide how big the box should be.
What if.. certain phrases exposed file attachment options in e-mail.
See kayak's guessing the origin of your flight by IP address.
Upcoming and Outside.in do this to show you local events.
What if... shipping costs were automatically calculated for me. eBay did this to just show you this information assuming the zip code on your account.
Bump most clicked items to the top of select lists. Abandon select lists and use autocomplete in text-input fields when the set becomes too big. If the app notices that you always look for the same things, pre-select these in the select list. This is an example of starting with a simple interface that adapts in tiny ways.
Designers love sliders but they aren't accessible. Provide text alternatives but if a user never uses them, hide them over time. You could also watch where the mouse is. If you miss a button a few times, make the button larger so your don't miss it the next time around.
The next version of Firefox will remember site specific text size settings. iPhone uses this to make the keys bigger as you type. BBC adjust coloring of content areas to track your common paths. Shaun Inman uses this on his blog for archive color. iGoogle changes throughout the day.
What if... we moved placement of text based on amount of text. Use a little information design, to emphasize the important stuff. If you do the same exercise a number of times you notice that the design changes in different contexts. The best way to display the same stuff changes depending on the situation. Use time based layouts to change emphasis. Highlight the stuff that matters.
Find: A List Apart article on contextual layout
Make transitions to a radically new release earlier for current users.
Google Maps groups the basic stuff closest to your house and lets you collapse it. You know how to get to the highway and Google affords that in its design. Google can track what you do and show you the options you always use and one more to promote your use of new features.
What if.. we changed help text based on audience. Let's book a trip v. Add trip details. Change phrase based and data.
What if... the navigation label varied from region to region. Coke, Pepsi, pop, etc. Change labels based on relevant data.
How much help text should you add? Why not... Adapt based on use. Shrink the help text after high frequency of use.
If we know that we are booking something for a particular situation. Replace navigation with what the person asked for.
Many applications are still based on the L shape. This is the wrong pattern for application design.
Instead use "where am I?"
Where I came from? What can I do next? This is how the iPhone works. The frame should recede and doesn't need to always be there.
What can we know about a person? Use referral URLs. If someone uses a number of Web 2.0 sites affordance becomes less important. For someone who doesn't use Web 2.0 sites add more discoverability tips for interface elements.
But is all of this possible?
Given the right data, all of this is easy. Cookies, openID, etc.
You can paint a detailed picture about someone from this data.
So a model for this.
There are three main sources of data:
- external (IP, etc.) Where I've been, what I've looked at, what I've done
- historical
- point-in-time
This is all about presenting information in the most desirable fashion.
We spend too much of our time looking for the silver bullet.
Caveats
Get the basic right before trying this. Communication and disclosure is vital to the success of this approach. Doing it wrong is worse than not doing it at all. Be careful with how you interpret data.
So how do you support this? Where technologies are moving with customer support where you can actually view a screen.
What about test-driven design? Turns out to not be as big a deal as you might think.
Keep it small. No BIG changes. Allow for easy profile switching to not muddy the data.
Be wary of messing with spatial information. When dealing with identity related data, treat this as a relationship.
So does this sound interesting or useful?
What if > Why Not > if/else
Look into ID 2.0 and game design to start down this path.
Check out poetpainter.com for more information.
0 TrackBacks
Listed below are links to blogs that reference this entry: AP UX Week '07: Creating the Adaptive Interface.
TrackBack URL for this entry: http://www.samfelder.com/mt/mt-tb.cgi/308
Who is this guy?
Leave a comment