Skip to main content
UX Design Engineering
Settings

Settings

Theme
Text size
100%
Typeface

Chapter 3

Defining success and choosing scope

Compare responses to a product problem, choose a feasible first release, and define evidence that could change the decision.

The team now has several plausible responses. It could add bulk assignment, clarify eligibility, improve recovery, or investigate the workflow further. Everyone agrees that administrators should finish their work more easily. That agreement is too broad to guide a release.

You need to choose the improvement to pursue, explain its cost, and establish what would count as progress.

Connect the user outcome to the business objective

For Casework, the user outcome is completing an appropriate assignment and knowing what happened. The business objective is reducing avoidable handling work while preserving correct routing. Faster submission is useful only when the work reaches an appropriate specialist without creating additional correction effort.

Write a chain that another person can challenge: clearer eligibility may reduce unnecessary attempts; fewer unnecessary attempts may reduce time spent resolving assignment failures. Each connection is a hypothesis until supported by evidence.

Avoid inventing a precise savings estimate from the supplied observations. We have not established how often the interruptions occur or how much of total handling time they explain.

Choose measures before choosing a target

Define what an attempt, a completion, and a failure mean. An attempt begins when the administrator submits an assignment. A confirmed assignment is one the service records successfully. A recoverable rejection is different from an unknown outcome after a lost response.

Useful measures for this case include successful task completion, time spent determining eligibility, repeated submissions on the same unchanged case, incorrect routing, and support requests about uncertain outcomes. Include keyboard completion and accessible recovery as release criteria, regardless of whether they move a headline metric.

If a baseline is unavailable, plan to establish one. Choose an initial cohort and a comparable task, record relevant context, and agree on who analyzes the result. A prototype task time and production handling time describe different environments; do not treat them as interchangeable measurements.

Compare complete options

The following comparison is a fictional planning exercise. Its costs are relative judgments that a real team would need to check.

OptionExpected benefitMain uncertaintyDependency
Bulk assignmentLess repetition for eligible groupsFrequency of safe groups and partial-failure behaviorNew service support and interaction work
Eligibility explanations and single-case recoveryBetter decisions before and after an attemptWhich reasons the service can expose usefullyExisting service plus limited reason information
Workflow guidance without interface changesFaster learning with low implementation effortWhether guidance compensates for missing system informationOperations support
Further investigation before a featureLess risk of investing in the wrong responseWhether delaying change leaves an urgent problem unresolvedResearch access and time

Compare maintenance and measurement effort too. A local workaround may appear cheaper until the product team accepts responsibility for keeping business rules in sync with a service it does not own.

Make a bounded recommendation

The worked recommendation is to test eligibility explanations, then deliver a dependable single-case assignment flow if the evaluation supports it. Bulk assignment remains an option, with grouping frequency and service behavior as open questions.

This is a staged investment. The first release should make current eligibility and outcomes clearer while retaining server authorization. It should not claim that information loaded earlier guarantees a later assignment will succeed.

Write exclusions directly: no selection across pages, no bulk mutation endpoint, no product-specific copy of authorization rules. Identify the decision that would reopen each exclusion. This helps a reviewer distinguish deliberate scope from a forgotten requirement.

Decide what would change your mind

Suppose administrators already understand eligibility and spend most of their time repeating otherwise straightforward submissions. That would weaken our first-release recommendation and strengthen the case for batching. Suppose the service cannot supply reliable eligibility information at all. We might narrow the interface to better outcome recovery while negotiating a service change.

Include these conditions in the decision log. They make it possible to revisit the scope without pretending the original decision was certain.

The product lead accepts the investment tradeoff; engineering confirms feasibility; the researcher or analyst reviews how the team will evaluate it. Until those conversations occur, your document is a recommendation.

Practice: choose the first release

Compare at least three responses, including an option that reduces implementation or delays a commitment. Record the evidence, assumptions, dependencies, maintenance costs, and exclusions for each.

Recommend a bounded release. Define the user outcome, business objective, measurement owner, baseline gap, and quality conditions. Give two observations that would cause you to revise the recommendation. Ask a peer to argue for a different option before you finalize it.

Review your work

The recommendation should be understandable without a mockup. A reviewer should see why you chose this scope, what you are deferring, and who accepts the consequences. Measures should have definitions and owners. Targets should be justified or explicitly marked as proposals awaiting a baseline.

The next chapter examines how to communicate eligibility and assignment outcomes through the interface.

Found an error? Report it.