Practice
Each case starts with a taught lesson and ends with an independent artifact. Return to a case when you want to test your judgment in a different context.
Open a case to study the method, get feedback on a guided decision, and produce the independent artifact.
20 cases
New context: A warehouse handheld scanner saves a stock count locally, then synchronizes when connectivity returns.
Make: A one-page task and handoff map with an annotated confirmation and recovery sequence.
The map connects an interface detail to the worker's actual outcome, includes the offline state, and does not claim remote success from local capture.New context: A physical parcel locker accepts a package and briefly lights a green indicator while its network is unavailable.
Make: An annotated three-state sequence: accepted locally, registered remotely, and registration failed.
Labels and feedback correspond to real state, the physical setting is considered, and a failed handoff has a recovery route.New context: A voice-only smart speaker guides a person through pairing a home device.
Make: A barrier and alternative map for the pairing journey, including a repeat and recovery path.
The learner identifies who is excluded, distinguishes the setting from personal ability, and provides an equivalent way to complete the essential task.New context: A shared physical access terminal lets a supervisor revoke a contractor's building access immediately.
Make: A concise risk and control map with an annotated action sequence.
The proposal matches controls to consequence, identifies authority and privacy, and does not treat a generic confirmation as sufficient evidence of understanding.New context: An enterprise employee abandons a multi-step access request after a manager calls them into a meeting.
Make: A structured observation sheet with a separate question list.
The learner preserves the task sequence and interruption, labels inferences, and asks non-leading questions about state and recovery.New context: A voice assistant's help command is used rarely. The team assumes people do not need help.
Make: A decision-question-to-evidence matrix with limitations.
Each method is tied to a question, measurement validity is checked, and the learner avoids turning absence of use into absence of need.New context: A support knowledge base on the web reports 'Nothing here' after an agent searches; a product filter, access permission, or index outage could explain it.
Make: A one-page cause map and problem statement with proposed checks.
The learner distinguishes empty data, no-match results, and retrieval failure, with targeted checks instead of a premature component choice.New context: Visitors to a public library website repeatedly call staff while trying to renew a membership.
Make: A one-paragraph problem statement and evidence/uncertainty ledger.
The statement is solution-neutral, tied to a user task and consequence, and explicit about the strength and limits of evidence.New context: A worker books a shared electric vehicle through mobile, then unlocks it with a physical terminal.
Make: A journey and task-flow diagram with at least one recovery branch and ownership notes.
The map includes cross-channel handoffs, honest state transitions, and a recovery route tied to the worker's goal.New context: A factory operator selects one machine from a large equipment list on a shared tablet, sometimes with gloves on.
Make: A two-option decision table and annotated interaction sketch.
Alternatives are both plausible, the chosen approach serves the actual task and environment, and the learner states evidence that could reverse the choice.New context: An administrator publishes a revised public policy; some comments remain unresolved and legal review may be pending.
Make: An annotated state model and short rationale for the review step.
The model covers authority, version changes, downstream consequence, and recovery beyond a generic confirmation modal.New context: A public-service website team needs to know whether people understand eligibility before starting a long application.
Make: A prototype brief with fidelity choices, scripted states, participant task, and limitation notes.
The prototype represents the behavior needed for the question and avoids claiming evidence about untested dimensions.New context: A web form proposes an inline summary of corrections before submission.
Make: A hypothesis and evidence table with task, expected effect, competing risk, and disconfirming observation.
The hypothesis is falsifiable, tied to a user task, and includes both intended benefit and plausible harm.New context: An enterprise analyst needs to export a report with sensitive fields removed.
Make: A one-page formative test plan with task scripts, observations, and study limitations.
The participant tasks match the decision, avoid naming controls, and the observations can reveal both success and misunderstanding.New context: A self-service kiosk test finds that two participants leave after a confirmation screen, while one asks staff whether the task succeeded.
Make: A finding card with observation, interpretation, impact, uncertainty, and next evidence.
The learner distinguishes observed behavior from cause, describes the study's limits, and chooses evidence that could change the conclusion.New context: A smartwatch accepts 'pause workout'; a fictional six-person test found three audio pauses instead of workout pauses, two noticed only after another minute, and four never looked at the display.
Make: A before-and-after interaction script plus a findings-to-change matrix and retest plan.
The revision follows specific observations, handles ambiguity without assuming a display, and the retest measures correction and added friction.New context: An enterprise expense app queues manager approvals during an outage; payroll closes that evening.
Make: A one-page staged decision memo with constraints, dependencies, owners, and release checks.
The memo connects an honest user state to operational ownership and identifies a condition that would change the release decision.New context: A hospital patient receives a telehealth appointment link, then may need an in-person test at a different site.
Make: A service blueprint fragment with touchpoints, backstage events, owners, and recovery messages.
The artifact identifies the source of truth at each handoff, accounts for access barriers, and tests the complete task across channels.New context: A mobile banking team considers showing transaction details before a recurring transfer is scheduled.
Make: A decision memo of at most one page with recommendation, alternatives, uncertainty, owner, and review trigger.
The memo ties its recommendation to evidence and consequence, treats alternatives fairly, and makes uncertainty and decision ownership explicit.New context: A fictional community health center offers class registration through web, phone staff, and an automated chat assistant. Classes have limited seats and eligibility rules.
Make: A compact case dossier: journey map, finding cards, state model, test plan, decision memo, and revision trigger.
The dossier distinguishes facts from hypotheses; aligns all channels to the actual enrollment state; handles eligibility, access, authority, and recovery; compares alternatives; and uses test tasks that can overturn the recommendation.