A shared bill that nobody can decompose
Eight teams share one warehouse. Spend has grown steadily for a year. Every team believes their usage is modest and each is right about their own share. Leadership wants the total reduced and has asked for a plan by the end of the month.
What you would do first
Answer before revealing anything. The value of the exercise is entirely in committing to a diagnosis you can be wrong about.
- 1Attribute before optimising: tag queries and jobs by dataset and consuming team, and produce a first report even if the coverage is partial.
- 2Decompose the total by driver — bytes scanned, work repeated, bytes shuffled, bytes retained — because each has a different remedy and a different owner.
- 3Find the patterns nobody chose: dashboards refreshing far more often than they are read, scheduled queries scanning without pruning, development environments pointed at production-scale data.
- 4Compare the growth curve against volume growth. Cost growing faster than data means the drivers are design decisions rather than success.
What is actually going on
The trap
The fix that looks right. Read it even if you got the answer — especially then.
Impose a proportional reduction target on every team. Each team makes its cheapest available cut, the teams with the most disciplined pipelines have the least slack and cut real capability, the actual drivers — a full rebuild and an over-refreshed dashboard — are untouched because nobody owns them, and the total returns within two quarters.