Your cloud bill is a design flaw, not a cost of doing business

When we audit a client's AWS or Azure account, we almost always find the same thing: the bill grew 5% a month for two years and nobody noticed, because it was nobody's job to notice. Cloud waste isn't a billing problem. It's an ownership problem.

Five fixes, ranked by effort

• Delete the zombies (1 day): unattached volumes, idle load balancers, forgotten staging environments running 24/7
• Right-size compute (1 week): most instances run under 20% utilization, so drop a size or two and nothing changes but the invoice
• Schedule non-production (1 week): dev and staging don't need to run nights and weekends, which is 65% of the hours in a week
• Reserved capacity (2 weeks of analysis): commit to your stable baseline for 30–60% savings on it
• Re-architect the hot path (1–2 months): the one service generating half your bill usually has a cheaper shape: queues, caching, or serverless

Make it stay fixed

One-off cleanups decay. The durable fix is a monthly cost review with a named owner, budget alerts that page someone, and cost as a line item in every architecture decision. We set this up as part of every DevOps engagement. Boring, and worth more than most features.