Microinteraction Testing Protocol

Description

A five-step test of one small interaction: elicit expectations first, observe unaided use, step through with think-aloud, ask what it should remember, then ask for the one fix.

Saffer adapts user testing to single microinteractions, where trigger, rules, feedback and loops are tightly coupled, so the prototype must be as high-fidelity as possible. Because the need for the feature is usually assumed, the test targets four things: the true goal (not a process step), which data matter, whether microcopy is needed and understood, and timing and flow. Goals and data come mainly from conversation, copy issues from observation. Expectations, prior experience and the one desired outcome are elicited before exposure, then unaided use is measured, then a think-aloud walk-through asks the participant to explain the rules, which exposes gaps between the designed and perceived model. The procedure is run in two rounds with revision between. Saffer warns that small behavioural differences need large samples: at least 20 users, ideally hundreds or thousands, for quantitative claims.

Attributes

Loss type
Latent
Data
Mixed
Effort
Days

Steps

  1. Before showing the prototype, ask how participants expect it to work, whether they have used anything similar, the one thing they want to accomplish and whether any single piece of information would make it unnecessary.
  2. Have them use the microinteraction unaided and capture quantitative data (success, time, errors) during or immediately after.
  3. Step through it with them thinking aloud and ask them to explain its rules; note discrepancies with the designed rules.
  4. Ask what they would want it to remember about them if they came back tomorrow.
  5. Ask for the one thing that should be fixed.
  6. Revise the design and repeat with a second set of participants.
  7. Treat quantitative differences as indicative unless you have at least 20, and preferably far more, users; avoid pausing mid-flow in ways that make people miss feedback.

Sources

  • Microinteractions (Saffer, 2013)

Also known as

  • Testing microinteractions
  • Expectation-first microinteraction test

Categories

  • User Research
  • Usability & Evaluation