A team revises a refund form, then measures completed valid requests, correction time, and support contacts while watching people enter an account in a task session.
Chapter 14 / 14
Measure whether the task became easier
Combine task evidence and performance data to judge a design change.
Your goalChoose one proposed interface change. Write a success definition with a denominator, a baseline task, a guardrail, and a question for a follow-up observation.
01 / Begin with the task
The principle
A design change becomes useful when it improves a meaningful outcome for people. Measurement begins with the task, a baseline, and a prediction about what the change should improve.
- Task success
- The proportion of attempted tasks that reach a defined meaningful outcome.
- Baseline
- A measure of the current experience collected in a consistent way for later comparison.
- Guardrail
- A related measure that helps detect a harmful side effect of an improvement.
Define success as a real completed task, such as a refund request submitted with a correct account, rather than a button click. Include the denominator: who attempted the task, through which route, and during which period. Pair completion and time measures with observed sessions, feedback, support contacts, and error patterns. The numbers locate a change; the observations explain the work people had to do.
Before an experiment, write the expected effect, the population, and the guardrails. Compare like tasks and segments across rounds. If a new form increases submissions while corrections and support calls also rise, investigate the quality of completion. Report uncertainty, missing cases, and the next research question alongside the headline result.
02 / Follow a method
Apply it in four steps
- Define a task and its meaningful success state.
- Collect a baseline with the same task and participant criteria.
- Choose one outcome measure, one guardrail, and qualitative observation.
- Compare the change, investigate differences, and decide the next test.
03 / Compare outcomes
Strong and weak execution
A team counts clicks on the refund button after launch and treats the increase as proof of a better refund journey. The count leaves completion quality and recovery unexplained.
Why the outcome changes
The strong evaluation connects an observed task to valid completion and watches for added recovery work. The weak evaluation measures interest in an action rather than the result people need. A repeated benchmark helps compare the same task over time.
04 / Use the method
Practice on a real screen
Choose one proposed interface change. Write a success definition with a denominator, a baseline task, a guardrail, and a question for a follow-up observation.
Review your work
- The measure describes a meaningful task outcome.
- The denominator, population, and period are explicit.
- Qualitative evidence and a guardrail help explain the result.
Keep your observation, interpretation, and proposed change together. Use the related patterns to inspect complete interaction states.