Post-Incident Reviews in the UK: the 72-Hour Breach Clock

A blameless post-incident review that also gives you the UK GDPR Article 33 breach record and the NIS 72-hour report, worked through a UK retailer's incident.

By · · 6 min read

Illustration for the Incident Postmortem Generator guide

United Kingdom edition. Written for UK teams: UK GDPR Article 33 (72 hours to the Information Commission, and a record of every breach), the NIS Regulations 2018 incident duties, and NCSC guidance on post-incident reviews. US edition

A UK post-incident review often has a second reader besides the engineering team. If personal data was exposed, UK GDPR Article 33 gives the controller 72 hours from becoming aware to tell the regulator and requires a written record of every personal data breach, reported or not. If your service falls under the Network and Information Systems Regulations 2018, a significant incident has its own 72-hour report. A blameless review with an accurate timeline is the cheapest way to produce both, and the NCSC's incident management guidance points out that careful records matter most when a regulator or court may later look at them.

This edition works through an incident at a UK online retailer with the Incident Postmortem Generator, and shows which parts of the review answer which legal duty. It reflects the legislation as published on legislation.gov.uk; your data protection officer or legal adviser decides how it applies to a particular breach.

A new name on the form: the Information Commission

Since 30 September 2026, the revised text of UK GDPR Article 33 names the Information Commission as the body to notify, replacing the Information Commissioner; the change was made by the Data (Use and Access) Act 2025 (Consequential Amendments and Transitional Provision) Regulations 2026 (S.I. 2026/386). The duty itself is unchanged:

Article 33What it means for the review
(1) Notify without undue delay and, where feasible, within 72 hours of becoming aware, unless the breach is unlikely to result in a risk to people's rights and freedoms; give reasons if laterThe timeline must show the moment the organisation became aware, and the notification time
(2) A processor tells the controller without undue delaySupplier timestamps belong in your timeline too
(3) The notification describes the breach (categories and approximate numbers of people and records), a contact point, likely consequences, and measures taken or proposedYour impact, root cause and action item sections, in plain language
(4) Information can be given in phasesNotify on time with what you know; the review fills in the rest
(5) Document every breach: the facts, its effects and the remedial action, so the regulator can check complianceThe finished review is that record, including for breaches you decide not to report

Paragraph (5) is the one engineering teams overlook. A breach judged unlikely to cause risk still needs a record, and a short post-incident review with the reasoning written down is exactly that record.

When the NIS Regulations apply too

The NIS Regulations 2018 cover operators of essential services (OES) and relevant digital service providers (RDSPs: online marketplaces, online search engines and cloud computing services that meet the definitions).

  • An OES must notify its designated competent authority in writing of any incident with a significant impact on the continuity of the essential service, without undue delay and no later than 72 hours after becoming aware (regulation 11). Significance turns on the number of users affected, the duration and the geographical area. The notice gives the time the incident occurred, its duration, its nature and impact, and any cross-border impact.
  • An RDSP must notify the Information Commission of an incident with a substantial impact on its service, also within 72 hours of first becoming aware (regulation 12).

Both lists of required details read like the summary block of a good review: when it started, how long it lasted, what it affected. Recording those as you go saves rebuilding them under deadline.

Worked example: order emails sent to the wrong customers

A UK retailer deploys a change to its order confirmation service on a Friday evening. A caching bug means some confirmation emails show another customer's name and delivery address. The review uses UTC throughout, because the email platform, the cloud logs and the support desk each record time differently, and says so in its first line.

In the Incident Postmortem Generator, enter one event per line:

18:32 UTC - Order confirmation release goes live; template cache shared between sessions
19:05 UTC - First customer emails support: "this is not my address"
19:40 UTC - On-call engineer confirms the defect; DPO informed (organisation aware: 72-hour clock starts)
19:58 UTC - Confirmation emails paused; exposure stops
23:10 UTC - Fix released; queued emails re-sent correctly

Enter 33 for time to detect, 86 for time to mitigate and 278 for time to resolve. The report shows 33 min, 1 h 26 min and 4 h 38 min.

What the review then records:

  • Scope. Email logs show 1,214 confirmations sent with another customer's name and delivery address between 18:32 and 19:58 UTC. That number goes into the impact section and, if the breach is reported, into the Article 33(3) notification as the approximate number of people and records.
  • The Article 33 clock. Awareness at 19:40 UTC on Friday gives a deadline of 19:40 UTC on Monday. The 72 hours run through the weekend, which is why the review meeting can wait until the following week but the risk assessment cannot. If the DPO decides the breach must be reported, the notification can go in on time with what is known and be completed in phases.
  • The record either way. If the DPO concludes the breach is unlikely to result in a risk, the review still documents the facts, the effects and the remedial action, as Article 33(5) requires, with the reasoning.
  • Detection. For 33 minutes, the only alarm was a customer. That is the first action item: "Alert when a confirmation email's address does not match the order's address, owner: checkout engineering lead, due in two weeks."

The trigger was the release. The root cause is that the template cache was keyed without the session, and no test checked one customer's data could not reach another. Action items follow: key the cache by order ID (solution), add a cross-session data test to the release pipeline (prevention), and add a one-click switch to pause outbound email (mitigation). The generator groups them, so a list that is all "fix the bug" stands out.

Running the review the NCSC way

The NCSC's guidance on cyber incident response processes divides a post-incident review into lessons from the incident (what would have prevented it or detected it sooner, tactically and across several incidents) and lessons from the response (what worked, what could have been better, and what data was missing, such as logs that were overwritten early). Built into a blameless review:

  • Ask what, not who. "What let a shared cache reach production?" produces action items; "who wrote the cache?" produces silence. People who expect blame leave out the timestamps a regulator will ask for.
  • Hold it within a week and share the draft first. Walk the timeline before discussing causes; most arguments are about when things happened.
  • Leave unknowns blank. The generator prints "Not recorded" for an empty duration. A gap you admit is better in a record the regulator may read than a guess that turns out wrong.
  • Track the actions to closure in the team's normal backlog, and check them at a regular meeting until they are done.

For a long chain of causes, the Root Cause Analysis tool helps with 5 Whys and fishbone diagrams; the Time Zone Converter helps put supplier logs on the review's clock.

Tools for this job

For the general method, see the blameless postmortem guide.

Frequently asked questions

When does the 72-hour UK GDPR clock start?

Article 33 counts 72 hours from when the controller becomes aware of the personal data breach, not from when it started. The review's timeline should show that moment explicitly, along with the time the notification was sent.

Do we have to record breaches we do not report?

Yes. Article 33(5) requires the controller to document any personal data breach, with the facts, its effects and the remedial action taken, so the regulator can check compliance. A post-incident review with the decision and reasoning is that record.

Who do we notify now the ICO has changed?

The revised Article 33 names the Information Commission, following amendments that took effect on 30 September 2026. The 72-hour duty and the content of the notification did not change.

Does the NIS 72-hour report apply to my company?

Only if you are an operator of an essential service or a relevant digital service provider (an online marketplace, online search engine or cloud computing service within the definitions). Then a significant or substantial incident must be notified within 72 hours of becoming aware.

Sources and further reading

Open the Incident Postmortem Generator

More Documentation & Review

Every tool in the Documentation & Review section, free and in your browser.