Cost optimisation reports identify savings that are never realised, because nobody owns the change and nobody verifies it landed. This tracker follows each opportunity to a confirmed reduction on the bill.
Start with waste, not with rightsizing. Orphaned disks, idle load balancers, unattached IPs, and forgotten non-prod environments are pure waste with no performance risk. Clear those before touching production instance sizes.
Record realised saving, not estimated. Vendor tools estimate optimistically. Compare the bill before and after and record what actually changed, or the programme reports savings that never appear.
Commit only to a stable baseline. Reserved instances and savings plans lock you in for one to three years. Commit to the floor of your usage, not the average, or you pay for capacity you stopped needing.
Here's a preview. Unlock the free download to get all 3 sections (2 more below).
| ID | Type | Resource | Finding | Est. monthly saving | Risk | Owner | Status | Actioned on | Realised saving |
|---|---|---|---|---|---|---|---|---|---|
| OPT-01 | Idle | 12 unattached disks in prod-001 | Detached over 90 days | £180 | None | Platform | Done | 2026-10-02 | £176 |
| OPT-02 | Idle | dev cluster running 24/7 | No activity outside 09:00-18:00 weekdays | £940 | None | Engineering | In progress |
More free templates IT teams use alongside this one.
Cloud cost allocation stalls on shared services: the network, logging, and identity that nobody owns and everybody uses. Deciding the split rule in advance is what makes showback credible rather than a monthly argument.
Untagged cloud resources make cost attribution guesswork and orphaned resources invisible. Tagging only works if it is enforced at creation. Retrofitting tags across thousands of resources is a project nobody finishes.
An IT budget presented as a list of line items invites line-by-line cuts. Split into run, grow, and transform, it becomes a conversation about what the business is choosing to fund, which is a far better conversation to be in.
Waste rather than rightsizing: unattached disks, idle load balancers, unused public IPs, old snapshots, and non-production environments running overnight and at weekends. These carry no performance risk and need no capacity analysis.
Only against a stable usage floor you are confident will persist for the commitment term. Commit to the baseline rather than the average. Paying for committed capacity you no longer use eliminates the saving entirely.
Because opportunities are identified in a report but never assigned an owner, and because estimated savings are reported rather than verified against the bill. Tracking each opportunity through to a confirmed reduction fixes both.
