Automated tree testing - faster, better, smarter?
Short Session, presented by Sam Ng.
What would we learn if we ran 300 people through a tree test of an IA? What if we then made on the fly changes and test another 300 people? Presenting case studies of projects run with automated tree testing software.
Getting stakeholders to agree an IA within a large organization can have its challenges. Opinions and subjectivity abound and it's hard to have a truly iterative approach to meet looming deadlines.
What if you could rapidly validate a site structure so that you could iterate more quickly? This session presents our learning from 4 projects that used automated tree testing to validate IA designs.
Tree testing is a method that gets objective data about the find-ability and labeling of items in a tree structure. It is like conventional usability testing, but simpler. Each participant attempts a series of tasks by clicking through a simplified site tree - drilling up and down - until they find a good match, or give up. Results show patterns in paths and whether people found the “right” category for each task.
By automating tree tests, hundreds of participants can be tested in days and results automatically generated so that iteration of designs can take place faster and with greater confidence.
Sam Ng
Sam trained as an industrial designer but quickly realized there were few industrial design jobs in NZ - so he co-founded NZ's leading UX consultancy Optimal Usability. At Optimal Usability, Sam has personally worked with more than 40 clients and on projects as diverse as retail customer experience, national park conservation through to online dating.
With 6 years self-taught experience in all things UX, web and business, Sam now leads a small startup team to create smart software for UX professionals. The first product, OptimalSort, is an online card sorting tool that has rapidly gained popularity – with over 110,000 end users in just over 12 months. A second product OptimalTree, is due to be released in late 2008.