UI Design for Non-Designers - From Models to Hypothesis to Metrics

January 8, 2010

Agree on a model

User interface design is filled with opinion and belief - something that many IT teams struggle to resolve (but enjoy discussing at length). A/B Testing adds science to the field, but is relatively late in the process of application development.

We need a better place to start when choosing an approach to application design. After all, our initial designs will set the style and standard of our application in users minds.

For our team, these three questions usually put us on the right path: * What is the end user trying to achieve and what is the simplest possible process to enable that? * What models or patterns can we copy from respected or inspirational sites to achieve that process / interface? * How can we make the experience consistent with our existing sites / applications? (If the new idea is better, how do we update them to match? If they are better, why are we doing something different now?)

The design hypothesis

All our applications go through a design review process during the Implementation phase of the Clarify, Simplify, Implement process.

First we discuss, challenge and agree on these UI models to be used. This is only be effective if the application designers have already been through these questions, asking them honestly and challenging their own design. (Don’t just look around for post-build justification.)

After agreement on these core principles and appropriate changes, the design review moves into a second phase of optimisation of the current design. This is where each field needs to be defended, pixel alignment becomes important and we do everything we can to reduce complexity and increase the aesthetic appeal of the application.

Optimisation with metrics

Now we have created our hypothesis of the best possible design for our application, we need to validate and improve it through A/B Testing, customer feedback and metrics.