$12K/mo azure spend recovered.

What we walked into.

Three Azure environments (prod, staging, dev/lab) were sharing a flat subscription model with no chargeback. Storage costs had ballooned because lifecycle policies were never enabled. Compute was running 24/7 in dev. Reserved instances had been bought once and never renewed. Finance had stopped trusting the bill, and engineering had stopped explaining it.

Week by week.

1. 1

   Week 1: Full spend audit — line-by-line cost mapping across all three environments, tagged by workload and business unit.

2. 2

   Week 2: Right-sizing pass — identified $4K/mo in oversized SKUs across compute and SQL.

3. 3

   Week 3: Reservation strategy — modeled 1-year vs 3-year RIs against 12 months of usage; recommended a $48K upfront commit returning $5.5K/mo savings.

4. 4

   Week 4: Lifecycle policies + idle-resource cleanup — moved 8TB of cool-tier-eligible blob storage and shut down 22 idle dev VMs.

5. 5

   Week 5: Azure Policy guardrails — deployed a baseline policy set so future drift gets caught at deploy time, not invoice time.

6. 6

   Week 6: Chargeback report template — finance now gets a monthly per-BU breakdown they trust.

What changed after handoff.