Lean UX asks a team to explain what a proposed feature is expected to change before treating its delivery as success. A finished dashboard, for example, says little about whether customers can make the decision that brought them there. That requires a clearer account of the problem and evidence about the experience.
Associated with Jeff Gothelf and Josh Seiden, the approach brings design, research, product, and engineering into a shared process of learning. Gothelf’s Lean UX Canvas connects business problems and assumptions to hypotheses and experiments. The practical value lies in making a proposed course of action open to examination.
Find the assumption carrying the decision
Consider an illustrative team planning automatic expense categorisation. It may assume that manual categorisation is a serious burden, that its suggestions will be accurate enough, and that employees will trust them. Those are different uncertainties, requiring different evidence.
User interviews can explore recent expense submissions and the work surrounding them. A prototype can reveal whether people understand a suggested category and can correct it. A technical evaluation is needed to assess classification accuracy. Calling all three “validation” hides the differences between what has been tested and what remains uncertain.
Choose the next activity by considering both the consequence of being wrong and the strength of the evidence already available. A cheap test of a minor detail offers little protection if the central assumption remains untouched.
Make the smallest useful test
Small does not mean carelessly made. A prototype test still needs suitable participants, believable content, and tasks that do not teach the answer. An experiment intended to measure a change in completion needs a defensible comparison and enough observations for the decision.
Write down what would lead the team to continue, revise, or abandon the idea before seeing the results. Those criteria can be qualitative or quantitative; a product hypothesis does not automatically call for a statistical significance test. Ballpark’s concept testing can help examine a proposed experience before a working implementation is available.
Keep the learning available to the team
Lean UX reduces documentation that serves no useful purpose, but some records carry real value. Future colleagues need to know which version was tested, who took part, what was observed, and why the team chose its next step. A concise decision record can preserve this without turning every study into a large report.
The approach also depends on room to change direction. If the feature and deadline are fixed regardless of the evidence, research has little influence over the outcome. Agreeing how findings will affect the plan is therefore part of the work, alongside designing the test itself.