The Six Parts of a Workable Privacy Request Process
None of these need expensive software. Most of the value is in the map and the runbook, with integration work targeted at the two or three places it genuinely saves effort or prevents a mistake.
System by system register Build the personal data map
A single table listing every system that holds customer personal information, what it holds (name, email, phone, addresses, order history, payment tokens, loyalty balance, support transcripts, notes), which system is the source for each field, which integrations copy it where, the search key you use to find a person, whether you can export, edit and delete through the admin screen, and who has access. For a typical retailer this is one or two days of work. It is also the document the rest of the process depends on, and it doubles as preparation for a data breach assessment.
Verify identity before releasing anything
Sending a customer file to the wrong person is itself a privacy incident. Match the verification to the sensitivity: for a correction to a marketing preference, confirming control of the email address on file may be enough. For a full access request that includes addresses, order history and support transcripts, ask for something only the customer would know, such as a recent order number and the delivery postcode, and send the response only to contact details already on the record. Do not collect a copy of a driver licence unless you genuinely need it, because then you are holding identity documents you also have to protect.
Search every system in one pass
Using the data map, a staff member searches each system by email, phone and customer number and exports what they find into a single response folder. Where the volume justifies it, an integrated search can query the systems with an API in one step and assemble a readable summary, leaving only the systems without one, such as the courier portal or the returns spreadsheet, for a manual check. The output should be in plain language: what you hold, where it came from, what it is used for, and who you share it with, not a raw database dump the customer cannot read.
Make corrections stick
The rule is simple: correct the field in the system the integration treats as the source for that field, then let the integration carry it outward, then check the downstream systems the next day. If the source is Shopify, editing the address in the ERP achieves nothing for long. Where a field is two way synced, confirm which side wins. APP 13 also expects you, if the customer asks, to take reasonable steps to tell other organisations you previously disclosed the information to, which can include your 3PL or a wholesale partner.
Delete, de-identify or retain Handle deletion with retention in mind
For each system, decide in advance what a deletion request means. Marketing profiles, loyalty accounts, reviews and support transcripts can often be deleted outright. Order and invoice records needed for tax usually cannot, but the customer name can sometimes be de-identified in the commerce platform while the financial record in the accounting system is kept. Suppression entries are a special case: keeping a hashed email on a do not contact list is what stops the address being marketed to again, so removing it entirely can work against the customer.
Log every request and outcome
A simple register: date received, channel, request type, identity check performed, systems searched, actions taken in each, date responded, any refusal and its reason, and who handled it. It lets you show what you did if the customer complains to the OAIC, it shows where the process is slow, and over a year it tells you whether volume justifies more automation. Keep the register minimal so it does not become another store of personal information.