As asked
The data platform bill has doubled in three months. No single team admits ownership. How do you find and control the cost without blocking useful work?
Sample answer outline
First make cost attributable by tagging warehouses, jobs, users, and datasets to teams or domains. Identify the largest spend drivers: long-running transforms, unbounded ad-hoc queries, over-sized clusters, excessive materialisations, and duplicated datasets. Add guardrails such as query limits, budgets, workload queues, auto-suspend, and review for expensive scheduled jobs. Partner with teams on optimisation instead of surprising them with quotas. Strong candidates pair cost reduction with platform ergonomics, because uncontrolled self-service usually means the platform is too opaque.
Expect these follow-ups
- Which cost cut would you avoid because it hurts productivity too much?
- How do you charge back or show back data platform usage?
- What query patterns usually cause the worst waste?
finopswarehousegovernance