Design a role-based personalized onboarding system

WHAT THIS TESTS: This question evaluates whether you treat onboarding as an extensible growth system rather than a static UI feature. Interviewers want to see separation of concerns between content, business rules, and presentation; multi-channel orchestration; and instrumentation for iterative improvement. They also look for awareness that onboarding spans multiple sessions and channels, not a single product tour.

A GOOD ANSWER COVERS: First, content abstraction. Propose a headless CMS, remote config, or structured database that stores onboarding content (tasks, tutorials, copy) outside the codebase so product managers can create or modify paths without engineering deploys. Each role maps to a content variant ID, not inline text. Second, a rule-evaluation layer. A lightweight service or edge function evaluates the user's self-reported role alongside behavioral signals (completed steps, time since signup, feature usage) to determine which path variant to serve. This keeps logic centralized and testable. Third, multi-channel delivery. The system should orchestrate in-app checklists and tooltips with behavioral email sequences and contextual help, because onboarding continues between sessions. Fourth, instrumentation. Fire events at every step (viewed, dismissed, completed) tied to the variant ID so you can measure activation rate per role and run A/B experiments on path changes. Fifth, progressive profiling. Acknowledge that role may change or that users should graduate into advanced flows later, so the architecture must support ongoing segmentation rather than a one-time assignment.

COMMON WRONG ANSWERS: Hardcoding role-based components in the frontend with if-else switches that require a full release to change. Treating onboarding as a single product tour that fires once on signup and never returns. Storing content in code repositories with no CMS or admin interface. Ignoring analytics and proposing to build the flow without defining activation metrics per segment. Failing to mention email or other out-of-product channels.

LIKELY FOLLOW-UPS: How would you add a new role without an engineering sprint? How do you handle users who skip the role selector? Can you run an A/B test on two variants for the same role? How do you prevent onboarding fatigue if a user sees too many prompts? How would you localize this for multiple languages?

ONE CONCRETE EXAMPLE: Imagine a JSON config where each role key points to an array of task objects with content IDs, completion events, and channel hints. An onboarding service reads the user's role, fetches the config, and publishes events to an in-app widget and an email provider. When the product team wants to add a Data Analyst path, they add a new config entry and upload copy in the CMS. No deploy is needed.

Source: appcues.com