BA ACADEMY / CASE STUDY
Digital onboarding: from drop-off to a testable plan
A fictional banking case showing how a BA moves from a conversion symptom to requirements and measures.
This is a fictional learning case. It demonstrates an analysis method, not a claimed client project or measured delivery result.
01
Business problem
35% of applicants leave before creating an account. The request is to add a progress bar, but the cause is not yet known.
02
Context
A retail bank offers mobile account opening with identity verification and a third-party document service.
03
Stakeholders
Applicants, Product Owner, operations, compliance, fraud, UX, engineering, the identity provider and support.
04
Discovery
Segment abandonment by step and device; interview recent applicants and support; observe failure paths; confirm regulatory and security constraints.
05
Analysis
Map AS-IS journeys, distinguish voluntary exit from technical failure, and compare error events with customer feedback. The biggest uncertainty is whether upload interruption or confusing consent language causes the drop.
06
Requirements
Support safe resumption after an interrupted upload; explain expiry and retry rules; record consent explicitly; capture event data for each step; protect sensitive documents.
07
Solution
Test a resumable flow and clearer guidance with a limited cohort before redesigning the whole journey.
08
Testing
UAT covers expired sessions, duplicate uploads, poor connectivity, alternate devices, provider downtime and accessibility.
09
KPIs
Completion within seven days, step-level drop-off, verification failure rate, support contacts and fraud-control exceptions.
10
Outcome
Reference outcome, not a real delivery claim: a decision-ready backlog and measurement plan that can be validated in an experiment.
11
Lessons learned
Do not treat a requested UI feature as the problem. Separate evidence, assumptions and risks; preserve controls while reducing friction.