Decommissioning AWS data you no longer need is one of the most reliable structural savings on a large estate, because storage that nobody owns only grows. The safe method is the opposite of a one time purge: find the candidates from billing data, assign each an owner to confirm, wait out a grace period, then tier or delete on a schedule. Done right, the waste stops accumulating instead of being cleaned up again next year.
This is an AWS storage, data, and network discipline. The instruments are simple; the risk is deleting something that still matters, so the playbook is built to make that nearly impossible.
Find what nobody is using
Start from the Cost and Usage Report and the storage consoles, not from intuition. The usual suspects on AWS are unattached EBS volumes left behind when an instance was terminated, snapshots whose parent volume no longer exists, old AMIs nobody boots, and S3 buckets with no recent access. Each shows up clearly once you reconcile the storage line against actual usage.
For every candidate, capture a suspected owner, a last accessed date, and a size. That metadata is what turns a scary delete into a routine one, because the decision stops being a guess and becomes a confirmation.
Assign an owner before you touch anything
The reason old data lingers is that no one is accountable for it. Tagging fixes that. Apply team, product, and environment tags so every storage object resolves to a person who can say keep or delete. Where a tag is missing, that absence is itself the finding: untagged storage is the storage most likely to be waste, and it is also the storage you must be most careful about deleting.
Notify the owner, give a defined grace period, and treat silence as a reason to investigate rather than a licence to delete. The grace period is cheap insurance against the one bucket that turns out to hold a compliance record.
Tier before you delete
Not every unused object should be deleted. Some data is rarely accessed but must be retained, and for that the right move is a cheaper tier rather than removal. The table below frames the decision.
| Situation | Action | Why |
|---|---|---|
| Unattached volume, no owner, no retention need | Snapshot once, then delete the volume | Keeps a recovery point while stopping the running charge |
| Rarely accessed but under retention | Move to a colder storage class via lifecycle rules | Retention met at a fraction of the cost |
| Snapshot with no parent and no compliance hold | Delete after grace period | Orphaned snapshots accrue cost for data nobody can restore to |
| Active but oversized backup set | Tighten the backup policy | Stops the set from growing without bound |
Make the cleanup a standing rule, not an event
A purge recovers money once. Lifecycle policies and ownership recover it permanently. Set S3 lifecycle rules to tier and expire objects automatically, schedule snapshot retention so orphans cannot pile up, and review the storage line on the same monthly cadence you use for the rest of the bill.
A Fortune 500 retailer carried years of orphaned EBS snapshots and abandoned buckets from completed projects. We tagged every object with a suspected owner and last accessed date, ran a confirmation cycle, tiered the data under retention, and deleted the rest behind a grace period. Lifecycle rules then kept the storage line flat. Figures are verified against billing data and anonymised.
Frequently asked questions
How do you find unused storage on AWS?
Is it safe to delete EBS snapshots?
How much can decommissioning data actually save?
Retire the data with us
We help enterprises find, confirm, and safely retire AWS storage nobody owns, then install the lifecycle rules and ownership that keep it from coming back. It is one chapter of the AWS cost optimization guide. Our guarantee: we reduce your cloud spend or we reimburse our service fee.
Put a defensible number on your cloud spend.
No provider in the room, no published price list. Tell us your footprint and we will scope the savings against your billing data — we reduce your cloud spend or we reimburse our service fee.
The Cloud Spend Navigator: what changed in cloud pricing, commitments, and FinOps — no vendor spin.