How much can we realistically reduce cloud costs by?

McKinsey's review of more than $3 billion in cloud spending found 10% to 20% in untapped savings at most organisations, including those with existing FinOps teams. Estates that have never been optimized, or that grew through rapid lift-and-shift migration, are usually at the higher end.

How do we reduce AWS or Azure cloud costs specifically?

The levers are broadly the same across providers: scheduling for non-production, right-sizing against telemetry, commitment discounts on a stable baseline, storage tiering, and control of data transfer. What differs is pricing detail and discount mechanics, so model each provider separately rather than assuming parity.

Does moving to the cloud reduce costs on its own?

Not by default. Cloud converts capital expenditure into operating expenditure and removes capacity planning risk, which changes the shape of the cost rather than the size of it. Savings come from elasticity, managed services and retired data centre overhead, and only where applications are built to take advantage of them. Our view on software solutions and a cloud-first strategy explains where the return actually comes from.

Who should own cloud cost reduction?

Accountability works best when it is shared: finance owns the forecast and the unit metrics, engineering owns the design decisions and the policies in the pipeline, and one executive sponsor owns the target. Programmes owned solely by finance tend to produce reports; programmes owned solely by engineering tend to lose priority to feature work.