Glossary. ROPA.
What is a record of processing activities (ROPA)?
A ROPA is the record of processing that GDPR and UK GDPR Article 30 require. It says why data is processed, so it has to agree with where the data really is.
By Dipal Panchal, co-founder and CEO · Updated November 8, 2026
Why data is processed, one activity at a time
A record of processing activities (ROPA) is the written record GDPR Article 30 requires of each processing activity: its purpose, the data and people involved, recipients, transfers, retention and security. Unlike a data inventory, which lists where data sits, a ROPA records why it is processed, so the two have to agree. A record written from interviews and never checked against the systems tends to describe the estate people remember.
This entry covers GDPR and UK GDPR, which share Article 30 in substance. India’s DPDP Act has no equivalent record.
One activity, field by field
The payroll activity of a UK company, Example Ltd, written against each point of Article 30(1).
| Article 30(1) | What it asks for | Payroll, as recorded |
|---|---|---|
| (a) | Controller, joint controller, representative and DPO contacts | Example Ltd, London; DPO at dpo@example.com |
| (b) | Purposes of the processing | Paying staff, and meeting tax and pension duties |
| (c) | Categories of data subjects and of personal data | Employees and leavers; name, address, bank account, salary, tax code, National Insurance number, sickness absence |
| (d) | Categories of recipients | HM Revenue and Customs, the pension provider, the payroll bureau (a processor) |
| (e) | Transfers to third countries, with safeguards where Article 49(1) is relied on | The bureau’s support team in India, under the UK International Data Transfer Addendum |
| (f) | Envisaged time limits for erasure, where possible | Six years after employment ends, per the retention schedule |
| (g) | General description of Article 32(1) security measures, where possible | Access limited to the payroll team; data encrypted at rest, with keys held by IT |
The same activity, as the systems hold it
One generic scan of the estate, not IRIS output, compared with the record above. The Why column says whether the record already covers each location.
| Source | Location | Detected type | Why | Sensitivity | Decision |
|---|---|---|---|---|---|
| Payroll system (SaaS) | employees.bank_details | Bank account, salary | In the record under (c) | Restricted | Matches |
| HR database | hr.absence.reason_text | Health: sickness reasons | In the record; the Article 9 condition isn’t written down | Restricted | Add the Article 9(2)(b) condition to the entry |
| Finance file share | finance/payroll/2025_export.xlsx | Bank account, salary, National Insurance number | Not in the record: a monthly export the payroll team saves | Restricted | Add the share to the entry, or stop the exports |
| Data warehouse | analytics.headcount_costs | Salary by employee ID | Not in the record: workforce analytics, a second purpose | Confidential | Record a separate activity with its own purpose |
| Payroll bureau, India | The bureau’s own system | Every payroll category | A recipient under (d); the bureau keeps its own Article 30(2) record | Restricted | Outside this scan; covered by the processor contract |
| Old payroll database | legacy_payroll | None | Still named in the record; decommissioned in 2024 and gone from the estate | None | Excluded from the scan; remove it from the entry |
What the two tables disagree on
Rows 3 and 4 are data the record doesn’t know about: an export saved every month, and a warehouse table serving a purpose nobody wrote down. Row 6 is the reverse, a system the record still names after it was switched off.
Copies don’t carry the purpose with them. The record names a purpose once, for the processing; the copies in the warehouse, analytics and CRM don’t say which purpose or consent covers them. Dipal Panchal, Sylox’s co-founder and CEO, has not seen a company record it against each copy.
Who has to keep one
Controllers keep a record under Article 30(1), and processors keep their own, shorter one under Article 30(2). Article 30(5) exempts organizations with fewer than 250 employees, unless the processing is likely to result in a risk to people, isn’t occasional, or includes special category or criminal offence data.
Payroll is never occasional, so even a small company keeps this entry. The record is in writing, electronic form included, and goes to the supervisory authority on request (Article 30(3) and (4)).
Check the record against a scan
Our rule: reconcile the record with a scan of the estate at least once a year, and whenever a system is added or retired. Two tests decide it: every system the record names holds the data it claims, and every system holding the data is named in the record.
Fix the record in both directions. Add what the scan found, and remove what it no longer finds, so the record a regulator asks for describes the estate as it is.
What a scan can’t fill in
A scan answers where. It can’t supply the purpose, the lawful basis, the recipients or the transfer safeguards, which come from the people who run the activity. Points (b), (d) and (e) stay with the business.
It also covers only the systems it reads. The bureau’s system in row 5 sits outside the scan, so the record leans on the processor contract and the bureau’s own record.
IRIS finds where each category of personal data sits in the sources you connect, so the Article 30 entry can be checked against the estate. Capabilities vary by edition.
Where each category of personal data sits: IRIS produces part of this, in the sources you connect.
Owner, purpose and retention beside each discovered table or file: IRIS produces part of this, as your stewards record them in the IRIS catalog.
The Article 30 entry itself: purposes, lawful basis, recipients and transfers: Not produced by IRIS.
Explore the data catalogReconcile one activity this quarter
Pick the activity with the most sensitive data, such as payroll or customer support, and list every system its entry names. Scan those systems and the shares and warehouses around them; the check passes when both tests above hold for that activity.
The record leans on a sensitive data inventory for the where. For the full comparison across activities, see the guide on Article 30 records against the data estate.
Terms to read next
What is a sensitive data inventory?
The step the record leans on: where each category of personal data really sits, by system and table.
What is data minimization?
The test the record makes possible: whether each field is needed for the purpose written down.
What is a DSAR?
The next test of the record: finding every place one person’s data sits when they ask.
Questions DPOs ask about the record
Article 30 doesn’t list it, but the UK regulator’s guidance recommends recording it, and you need it for the privacy notice under Article 13 anyway.
Yes. Article 30 asks for a record in writing, including electronic form, and sets no format. What matters is that it is complete and current.
Detailed enough to find the data again. “HR system” is too vague for a reconciliation; the system and the table or share are enough.
Protect your brand. Start with your data.
Bring one entry from your record. We’ll show you where IRIS finds that activity’s data across your sources.