1. Establish the operating context.
Identify system owners, important tasks, demand patterns and recovery needs. List known constraints and current access.
Process
Investigate first, then make a measured change.
The review begins with an agreed read-only scope and the information needed to understand the workload.
2. Connect usage and spend.
Compare the relevant time periods and resource categories. Note missing attribution and changes in business activity before proposing a cause.
3. Compare practical options.
Document the expected benefit, effort, assumptions and possible effect on performance or recovery. Keep unverified savings separate from measured results.
4. Test a limited change.
Use a suitable controlled environment or agreed release window. Check capacity and important tasks, with conditions for stopping or restoring.
5. Review the result.
Compare equivalent usage and operating conditions where possible. Record the actual outcome, limitations and owner of the changed setup.