UX London: Prototype & Test In Five Days

by May 28, 2014

In their How to Prototype and Test Any Product in Five Days presentation at UX London 2014 Daniel Burka and Jake Knapp shared how to vet and build software faster. Here are my notes from their talk:

  • Most software is created by build, learn, and iterate. This sounds great on paper but it rarely works out in practice. It usually takes more effort than you think to build something, at which point you need to launch it because its built, some people love it so you decide to keep it alive, and you are on to the next shiny thing, while supporting the old features.
  • How can you shortcut that process? Google Ventures has worked out a five day design process to work through ideas quickly.
  • Ship early and often forces you to meet deadlines. The five day design process starts by scheduling five people for interviews on a Friday. This adds a time crunch/timeline to the work process.
  • Include as many people as possible in the early phases (day one) of the process: design, development, marketing, founders, etc. Their perspectives will provide insights you can't get otherwise.
  • From there, explore a variety of solutions -without group brainstorming. Instead, work individually on ideas then present sketches together at the end of the day.
  • Talking is not an efficient way to come to a decision. Timebox the amount of discussion and use weighted voted to avoid design by committee. From there, talk about how the company makes design decisions -that should not be democratic. Design needs an opinion, so some people get extra voting rights.
  • Once you have a few ideas, how do you decide what to prototype? Can you test all the ideas quickly?
  • Make your initial sketches detailed from the start. Then move to a level of "medium fidelity" so people think they're looking at a site/app. Keynote is a great tool for quickly prototyping ideas.
  • Distribute the work. Get some people writing copy and many people working in Keynote (which is usable by more than designers). This allows you to move fast.
  • Aim for a fidelity that gets people to suspend belief -so they feel the app/site is real. Do research on this level of prototype.
  • A lot of companies don't want to do user research because they worry about needing time, experts, equipment, and more. But you don't need any of that just put people in front of a laptop or mobile device and talk to them.
  • Test with five people and compare against original designs (if available). This gives you a lot of data about WHY things work.
  • Early data like this provides enough confidence for companies to go build things for real.
  • Does this process work for expert users? Google Ventures has used the process for a variety of domains and found the fidelity works for them as well.
  • Create a deadline: you can get things done fast if you have to. Prototype quickly and do research.