How do you handle a one-off token exception for a campaign?

What's really being asked

This question evaluates whether you understand design system governance as a socio-technical contract rather than a style guide. Interviewers want to see that you can balance business velocity against systemic consistency, prevent design drift, and manage organizational politics without becoming a bottleneck or a pushover.

A GOOD ANSWER COVERS four things in order. First, governance: propose a lightweight fast-track for exceptions that still requires written justification, stakeholder sign-off, and a mandatory sunset clause such as thirty to ninety days post-campaign. Document every exception in a public drift log or exception registry so the team can spot patterns. Second, technical isolation: never mutate core primitive or semantic tokens; instead expose the one-off through a campaign-scoped override, an experimental token namespace, or a feature flag. Use a conspicuous naming convention like campaign-holiday-2024-neon so the hack is searchable and greppable. Third, long-term risks: visual fragmentation across surfaces, token bloat that slows build times, broken theming or dark mode, accessibility violations, and erosion of trust that teaches product teams to bypass the system. Fourth, communication: announce the exception to the design system channel, update documentation, and schedule the cleanup task in the roadmap before the campaign launches.

COMMON WRONG ANSWERS include refusing all exceptions on principle, which signals political inflexibility and a lack of business partnership. Another red flag is allowing a hardcoded hex value directly in a component file without any token abstraction or audit trail. Candidates also stumble by proposing technical workarounds but omitting a sunset clause, removal ticket, or drift log, which guarantees the exception becomes permanent technical debt. Ignoring the power dynamics entirely, such as letting an executive override bypass documentation, is also a failure mode.

LIKELY FOLLOW-UPS the interviewer may ask next include how you would respond if an executive mandated the change without process, what you would do if the campaign color fails WCAG contrast against existing backgrounds, how you measure drift or audit exceptions quarterly, and at what threshold of repeated exceptions you would recommend expanding the core token set instead of continuing with one-offs.

ONE CONCRETE EXAMPLE is a marketing team requesting a neon green not in the brand palette for a holiday banner. Governance step is filing a one-page exception request with a hard sunset date of campaign end plus two weeks, approved by both product and design system leads. Technical step is creating a campaign holiday2024 promo banner background token that lives only inside a campaign feature flag, leaving core colors untouched. The token is greppable, linked to a Jira cleanup ticket, and logged in the drift registry. If marketing asks for the same color six months later, the log proves the palette needs an official extension rather than another exception.