Guides

Is the Deletion Really Done? A DSR Verification Checklist

November 3, 2026

The deletion job finished. The customer record disappeared from the main application. A confirmation email is ready to send.

Then someone asks about the warehouse copy, the support attachment, and the vendor that received the record last week. The request is marked complete, but the evidence covers only one system.

Data deletion verification is the check between executing an erasure action and claiming that the request has been fulfilled. It connects the approved scope, the result in each relevant location, any justified retention, and the explanation sent to the requester.

For privacy teams, the useful question is specific: what evidence supports each statement in our response?

Why completion deserves its own review

The European Data Protection Board's report on implementation of the right to erasure, adopted on February 10, 2026, identifies recurring difficulties with retention periods, backups, anonymisation, and information provided to individuals. Its recommendations include verifying that erasure occurred and being able to demonstrate it. These recommendations are non-binding; the report does not introduce a new legal deadline.

The checklist below is a suggested operating process for privacy, IT, and vendor owners. Use it after confirming identity, the applicable rights, and the request's scope. Your legal team should determine which obligations and exceptions apply to the particular request.

1. Define what this request covers

Start with a compact scope record. Give the request an internal reference and identify the accounts, identifiers, business units, and processing activities it covers. Resolve ambiguous account matches before a destructive action.

Ask system owners to trace the relevant information through active applications, downstream stores, files, vendors, and backups. Include derived records that remain linked to the individual. A search by email alone may miss a record stored under a customer ID.

Mine's data mapping best practices provide a starting point for maintaining that inventory. For this review, turn the map into a list of locations that need a documented outcome.

Give each location an owner, a matching method, an intended action, and an evidence reference. Keep “not yet checked” visible. An empty result from one search should not silently remove a system from scope.

2. Separate execution from verification

For each system, record the operation that ran and the evidence that the intended result followed. A successful submission to a deletion queue establishes that work was accepted. Check whether the system provides a separate completion event, result report, or verification query.

Agree on the meaning of each outcome before automating closure:

Keep backup treatment in a separate field so it cannot disappear behind a successful live-system result. These are proposed workflow labels, not statutory categories.

Choose checks appropriate to the system. A post-action lookup, a scoped completion report, or a responsible owner's recorded verification may be useful. A screenshot of an empty user interface is weak evidence if the underlying records remain accessible elsewhere.

3. Get a scoped answer from each vendor

Send the vendor enough information to identify the request securely, then ask what its confirmation actually covers. Record the relevant service or tenant, the requested action, the completion date, any excluded data, and the treatment of downstream providers and backups.

California's Attorney General explains the CCPA right to delete as including requests that a business tell its service providers to delete, subject to exceptions. Determine the applicable notification duties for your request; sending an instruction and receiving evidence of completion are separate operational events.

If a vendor replies “completed” without a scope, ask for the missing detail. Assign a follow-up owner and escalation date. Keep a vendor's unresolved task from being converted automatically into a claim that every recipient has deleted the data.

4. Make retained data and backups explicit

For a proposed retention exception, ask Legal to identify the particular records, applicable basis, permitted use, and retention period or review event. Ask the system owner to document how that limited use is enforced. A requirement to retain an invoice should not automatically become approval to retain an entire marketing profile.

For backups, record which copies may still contain the information, how access is controlled, when those copies expire, and what happens if a restore occurs. Assign responsibility for carrying the erasure decision into the restoration process.

The EDPB report discusses the need to track erasure requests and address them on restored systems where backup integrity limits direct modification. It also identifies weak masking or pseudonymisation being mistaken for anonymisation. Treat backup schedules and transformation claims as matters to verify, not automatic evidence of completed erasure.

Have Legal and IT review the proposed backup handling for the applicable jurisdiction and architecture. A technical constraint should produce a documented decision and follow-up, rather than an unqualified completion message.

5. Test whether a routine data flow can undo the result

A deletion can appear successful until the next import recreates the record. Review recurring flows such as CRM-to-marketing syncs, warehouse refreshes, file imports, and recovery jobs. Identify which source could reintroduce the same information.

Test the workflow with synthetic records in an isolated environment. Run the deletion process, execute a representative sync, and check the downstream result. Test restoration controls separately with the infrastructure owner.

If preventing reintroduction requires a suppression record, define the minimum identifier needed, the permitted purpose, access controls, and retention rule. Review its legal basis. Avoid rebuilding the deleted profile inside the evidence or suppression store.

A worked example: four systems, four different outcomes

Consider a hypothetical retailer handling a verified deletion request. The customer account, marketing platform, support system, and finance archive each need a separate decision.

The useful result is a clear account of what happened in each location. In this example, the marketing failure prevents a blanket statement that deletion is complete. The invoice exception also needs an accurate explanation rather than being counted as a successful deletion.

6. Match the response to the evidence

Before sending the response, have the request owner compare it with the scope record. Describe what was done, what remains and why, and any relevant follow-up. Apply the required response timing and information requirements for the jurisdiction; an internal verification process does not create extra time.

Use a short closure check:

Protect the evidence itself. Prefer request references, timestamps, decision records, and restricted links over copies of the personal data being erased. Define access and retention for this record too.

Put verification into the DSR workflow

A completion check works best when it is part of the process from the start. Mine's guide to building a DSR workflow connects system identification, identifiers, downstream flows, and customer notifications.

MineOS DSR Automation supports customizable workflows, cross-system request handling, and detailed request logs. Use those capabilities to organize the process, and agree with each system owner what evidence will support its outcome.

Start with one recent deletion request. Trace every statement in its final response back to a result, decision, or vendor confirmation. Turn any missing link into a named task. Talk with Mine about bringing that discipline into your DSR operations.

Ready to build your own autonomous kingdom?

Book a demo

Ready to build your own autonomous kingdom?

Book a demo