Aaron Forth, VP of product for personal finance and budgeting service at Mint. com, talked at SXSW this week about the challenges the company faced when going mobile. As a service whose business model depends on people visiting the Mint.com website, the decision to launch a mobile application was not one the company took lightly, Forth said.

But Mint did decide go mobile, launching first on iPhone, then on Android, after studying the actions of its mobile user base. And now, Mint is planning to launch a brand-new iPad application in just a few months - a prospect that has the company again rethinking its product for the tablet's multitouch interface.

Case Study: Mint Goes Mobile

Mint says that the majority of its users have a phone connected their account with the service, and 20% are relying only on a mobile phone. Given those demographics, the company knew it wanted to serve the mobile audience with native applications.

Forth then explained the thought process the company went through when creating its first mobile app (the iPhone version of Mint). The results make for an interesting case study which other developers may be able to learn from.

Step 1: Pick Your Platforms Carefully

Forth said that when it came to making a decision about what devices to launch its applications on, the first thing they looked at was device adoption trends. What devices have good market share?, they wanted to know. The answer was Android and iPhone and Android, each which have "massive distribution," said Forth.

Another important factor was whether or not the platforms they chose would have a way to get apps directly to consumers via an app store environment. Again, both iPhone and Android filled this need.

Finally, they needed a platform where there was a good SDK  available and developer support. In this area, Forth said Mint found that Apple's strict guidelines didn't interfere with development, but actually helped them understand very simply what they could and could not do. They didn't have to figure out the phone, he said, they could just focus on development.

Another key factor in the decision was finding a platform where there was low fragmentation. "It's hard enough to support multiple browsers and versions," said Forth. Mint wanted to find platforms that had a fragmentation strategy plan. Apple is great example of this, he said.

Once the company decided what platforms it wanted to build apps for, it created some general goals. For example, the company decided it wanted to earn a top three position in the free finance category and it wanted to take advantage of specific platform features available to native apps, because that would make it more likely that platform maker would highlight the app within its store as a showcase example of its platform.

Finally, Mint wanted to deliver on-the-go utility to its users, and make Mint easy to access. But most importantly, because Mint needed visitors to its website to earn revenue, it decided to build its mobile app as more of a companion to the site, instead of a standalone application.

Step 2: Figure Out What's Transferrable

The second question Mint needed to address was to figure out what's reused versus what's new when going from the Web app to mobile. "You basically tear down the idea of what your product is when you go mobile," said Forth.

On the "reused" side, there was only brand and positioning, the finance data and the user profile that was ported to the mobile app from the Web. Everything else was new. This  included, of course, the UI itself - the interaction and visual design involving gesturing, a one-button interface on iPhone and the fact that there's not a lot of typing involved when using mobile apps. It also included new security features where encrypted protocols were used to exchange data, plus the use of passcodes and more.

Also new were to the company were the skills involved in the app's creation (languages like Objective C and Cocoa), the different development environments used, the QA process and the authentication methods.

Even the service layer and architecture was new, because the layer of APIs used for the earlier Mint Web app were different from ones that support the mobile app. On a phone, explained Forth, the app needs to be more efficient when delivering the payload of data to the device. It's bursting data to you upon loading, instead of streaming data to you, like on the Web.

Step 3: What is the End User Experience You Want?

Finally, Mint wanted to focus on the user experience, keeping in mind that the mobile app will serve different needs than Web app, and its design needed to evolve in order to embrace device it runs on.

Mint designed the app to focus on core jobs using simple screens: an overview screen, accounts screen, expenditures screen and budget screen.

In a later version of app, Mint made it easier for new users to sign up for the service directly from the app itself. This was important to the company because Mint's users with mobile phone are more engaged than the others. After 4 weeks, 20% of Mint's new sign-ups were from mobile, said Forth.

Mint Coming to iPad!

Now Mint is developing its first iPad application, which again is forcing the company to redefine what its app is. The tablet form factor offers new possibilities, and new use cases, said Forth. Within the iPad app, some of its graphic libraries will be built in HTML5, for example, when previously on the Web, its graphics were built with Flex.

It will also incorporate gesturing for things like categorizing transactions, which will work via drag-and-drop. Like its iPhone counterpart, the iPad app will be more similar to the iPhone than website, as there won't be a deep, exhaustive set of menus. It will serve as a companion to the website as well.

In the Q&A session at the end of the talk, an audience member asked why Mint didn't launch on iPhone and Android at the same time. "In an ideal world, we would," said Forth. But there were time constraints. Also, Mint's developers were more comfortable and familiar with iOS, which made it easier to launch there first. If there is etiquette around when to launch on different platforms, we probably broke it, Forth said.