|O Danny Blog Entries|
SXSW 2007: Mobile Application Design Challenges and Tips
Panel: Kevin Cheng (Yahoo!), Matt Jones (Nokia), Simon King (Yahoo! Research Berkeley), John Poisson (Tiny Pictures Inc/RADAR.NET), Anita Wilhelm (Caterpillar Mobile).
MJ: User research and prototyping for mobile is incredibly important and difficult. The mobile experience is difference when it is in your hand. The sooner you can get a sketch on a block of wood or something in people's hands is the most important. Post-it notes on blocks of wood. Some Flash Lite code: XML representation of screen flow. People react very differently when it is in their hand.
AW: The most successful method I've had is paper. In some cases with foamcore so people could push buttons. Diary study with a brick to see what they would want to do with that brick.
KC: How is mobile design different than web design?
SK: Much more fragmented and divided attention situations. 20 seconds at time.
KC: What sort of metrics do you use?
JP: Clicks are gold in mobile space. Saving clicks makes sense. We even used .net because it was easier than .com
SK: We do a lot with click tracking too. Number of clicks are used to fine tune. It is harder to maintain state: no deeply nested menus or breadcrumbs.
AW: Session time is important: 3-5 seconds is successful. Designing for mobile is about breaking up an application into small steps. Keep minds available for whatever else they are doing.
KC: How do you deal with interruptions?
AW: Graceful failure. Design for interruptions so they can be resumed.
JP: Very different than designing for desktop. Very discreet tasks. Never going to go through a long string of actions.
MJ: Always know where your default cursor position is. Don't put anything complicated in a dialog! Always give people a positive path through the application using the defaults.
Q: Who is actually using mobile applications? I don't believe many people know you can send a text message via email.
JP: That's true. But time and again, even people who have a minimal exposure to phone functionality can learn it easily. Needs to be explained to users in the right orders and the right sequence. And once they learn the pattern they can do it.
Q: You have all these given inputs and outputs on mobile. If you are thinking about solving a user goal, what makes you choose one or the other?
SK: Depends on what you need to do. SMS can be limiting. It depends on the problem you have to solve. One issue is integrating with the other applications on the phone. Need to consider target platform.
JP: Need to solve a particular problem.
AW: Need to consider the experience. You can customize outputs for that application. If you can put it all into one centralized location, it makes it more of one experience.
SK: This also makes it faster.
MJ: Look at a service design approach to it. Magic tablets and magic clouds that follow them around. Need to look in a human-centered way about the service.
KC: At Nokia, how much of your time is spent designing the entire experience around the phone?
MJ: More and more. Particularly in the area I work in. The interaction between the cloud and the little mobile thing. Not just about going through this tiny post-it note. How we decide what to expose when. Two most exciting things is mobile version of Apache server--makes the relationship different. Python for S60 is also very interesting for rapidly making new web services. A wonderful pattern is emerging as mobile as stubmaker. "This is very interesting to me" then going back later and filling in that stub. You can do that with SMS or whatever.
Q: Flash Lite isn't supported on most phones. Emulators aren't great. How do you test an app on a million different phones?
SK: Excellent prototyping tool, but not good for actual applications.
JK: It simply isn't the case you can write one and use anywhere. Alternative is put it out there.
SK: Pay someone to test it for you.
AW: Number of companies that will test online.
MJ: If you test the edge cases, you are 70% good.
AW: Certain groups of devices you can test one or two from a group.
JP: Make it work on one or two categories very well.
Q: What is your strategy for getting your applications used and on people's handsets? Do you go on the carrier on-deck or off-deck?
AW: Have both off-deck and on-deck distribution for scannr. Decision is in the business deals and monitesation. On-deck has more distribution channels and easier billing for services. But it is a pain. Off-deck is much easier and lets you control the experience better.
JP: Some different strategies for distribution. We used the WAP experience, especially for outside the US. Distributing the app via WAP site.
KC: Off-deck means you also don't have to negotiate with all these carriers, right?
JP: Yes. Put it out there and build demand.
AW: Ecosystem of mobile is much harder than web. You need a tiered system: give users what they can use. Multiple versions: free version, deck version, etc.
Q: What do think about browser development in the phones?
SK: Opera browser rules. Makes WAP much less cool.
JP: Drinking the ocean with a spoon.
MJ: Need to think about the partial attention.
KC: Need two different designs?
MJ: Need a sweet spot between the two. Most websites are pairing themselves down anyway.
Q: Do you have any favorite resources for design patterns? Books, etc.
MJ: Sound and haptic feedback are also important, by the way.
SK: There aren't any.
Q: Strategies for open access to things on the phone?
MJ: We're actually trying to open up more. It's becoming baked in to things like the S60.
SK: Another company is to keep an eye on is Open Moko an open hardware and software platform.
JP: Fair amount of pushback from carriers against openness.
MJ: Carriers in Europe and Asia are understanding more open = more money. Becoming "The Proud Pipe" with only a little added value.
Q: Is there a standard email SMS back-end?
AW: Can't send an SMS to an email in the states.
JP: Can't programatically send out SMSs to the phone. And the answer to the question is no. It's a never-ending battle.
AW: Startups are trying to go around the carriers, but carriers are pushing back.
KC: We're out of time...