OPERATIONAL RESILIENCE · CAREER SKILLS
FCA operational incident reporting: skills for payments teams
Operational incidents create pressure to explain impact before every fact is known. The new UK framework rewards disciplined work: clear thresholds, reliable evidence, named decision owners and updates that separate confirmed facts from assumptions. This guide turns the rules into practical skills for UK payments operations, risk, compliance and customer teams.
By Awais Amir · Published 2026-10-31 · Updated 2026-11-15
What changed, and what did not
The FCA finalised the rules in March 2026. Its implementation page was updated on 30 October 2026 and points firms to the policy statement, final guidance and reporting templates ahead of the 18 March 2027 start date.
The framework creates a standardised process across the FCA, Prudential Regulation Authority and Bank of England. A qualifying incident will be submitted through FCA Connect using one form. Most FCA solo-regulated firms will use a standard report. A smaller group of enhanced reporting firms will provide more information when the status changes and when the incident is closed.
The framework does not make every service interruption reportable. Firms must assess the facts against the regulatory thresholds and their own circumstances. Existing duties to communicate with supervisors may still apply.
Learn the three reporting thresholds
An operational incident reaches the FCA threshold when a firm reasonably believes it poses a risk in at least one of three areas: intolerable consumer harm from which consumers cannot easily recover; the safety and soundness of the firm or other market participants; or market stability, market integrity or confidence in the UK financial system.
The FCA says firms should consider factors such as the impact on end users, legal and regulatory duties, the ability to provide adequate services, data availability and integrity, and the firm's internal incident classification. This is not a tick-box list. Context and evidence matter.
- Consumer impact: who is affected, what they cannot do and whether recovery is easy.
- Operational impact: which services, systems, data and third parties are involved.
- Financial-system impact: whether the event could affect other firms, markets or confidence.
- Evidence quality: what is confirmed, what is estimated and who owns each assessment.
Use one evidence log from detection to closure
Strong incident work starts with one reliable chronology. Record the first observed symptom, affected service, geography, customer group, transaction stage and source of each fact. Add the decision owner and the time of every material change.
Keep symptoms, impact and cause separate. Failed card authorisations are a symptom. The number of affected users and blocked journeys describes impact. A supplier fault, software change or cyber event is a cause only when evidence supports it.
The PRA version of the framework uses one report that is updated through the initial, intermediate and final phases. That makes change control a core career skill. Each update should explain what changed, why the assessment changed and which facts remain uncertain.
Live evidence example: Microsoft 365 outage, 13 November 2026
At 09:58 GMT on 13 November, PXP's official status page marked Settlement as a major outage. Its notice said the Microsoft 365 disruption may affect daily settlement processing, chargeback reports and fraud notification services. The same page showed Transaction Processing and its token vault as operational. 'May affect' describes a current risk, not proof that every settlement failed.
Worldpay's status pages described a different impact. Worldpay for Platforms said email and scheduled transaction reports were unavailable, while settlements continued as usual. Worldpay's main page showed authorisation and payments, acquiring, and funding and payouts as operational, with portals and reporting degraded.
Independent reporting recorded Microsoft's statement that it had no indication of malicious activity. Keep the event classified as an operational disruption unless later evidence establishes a cyber cause. The practical lesson is to maintain a provider-by-provider component matrix and never transfer one provider's symptoms or severity to another.
Recovery check, 20:20 GMT on 13 November: Microsoft said it had addressed the underlying issue and was seeing broad recovery and stability, while restoration continued for remaining customers and regions. Microsoft's public status page showed all products operational when checked. PXP, Worldpay for Platforms, Worldpay and Cybersource still had their downstream notices open. Do not close a supplier incident solely because the upstream platform has recovered; wait for the affected provider to confirm its own service, reports and settlement position.
Provider closure check, 21:47 GMT on 13 November: PXP marked its incident resolved, recorded an end time of 13:30 GMT and confirmed merchant settlement was performed on time. Worldpay said email had largely recovered and confirmed it had seen no client-funding or transaction-processing impact, although its incident remained open for residual monitoring. This converts PXP's earlier potential settlement risk into a confirmed on-time outcome without changing the separate Worldpay or Cybersource records.
Worldpay closure check, 06:58 GMT on 14 November: Worldpay marked its main email incident resolved and said its ability to send and receive email had fully recovered. It repeated that client funding and transaction processing were not affected. Worldpay for Platforms still showed its separate report-delivery notice in monitoring when checked, so that record remains open until the provider publishes its own closure.
- Dependency: record the supplier named by the provider, without assuming other providers share the same cause.
- Component: separate settlement, reporting, support and transaction processing.
- Evidence status: label each effect as potential, observed, confirmed or resolved.
- Operational control: record the fallback contact route, report retrieval method and reconciliation owner.
- Next check: capture the provider timestamp and the time your team will review it again.
UK bank example: separate supplier incidents from maintenance
At 14:05 GMT on 14 November, Royal Bank of Scotland's business status page said the external Microsoft outage was affecting some NatWest Group services. The same page separately listed planned Bankline maintenance and routine maintenance affecting access to digital and mobile banking mailboxes and statements. It did not attribute those maintenance windows to Microsoft. Reuters reported that UK banks activated incident teams while most customer services remained unaffected. Record the supplier incident and each planned maintenance window as separate events unless evidence links them.
Post-incident review example: Revolut and Google Cloud
Revolut published a technical review on 13 November of its 20 October outage. It said a power failure affected one Google Cloud availability zone in London at about 06:00 UTC. Many customers could not make payments, withdraw cash or access the app. Backup processing allowed some card payments to continue. Card authorisations and in-store payments were restored by 12:40 UTC, app access and transfers began returning from 15:45, and all remaining services were back online by 17:15.
The external power loss was the trigger, but Revolut's review identifies separate internal resilience failures. One member was already missing from some five-server coordination groups and monitoring had not alerted the team. When two more members went offline, the remaining servers could not form the quorum needed to promote a database replica. Some recovery tools also depended on the affected zone, which slowed diagnosis and remediation.
A joint Bank of England, PRA and FCA letter dated 4 November corroborates the Google zone power loss and disruption to UK financial firms. It says no malicious activity was identified, no material ongoing consumer effects were found and there were no identified implications for market integrity or wider financial stability. That regulator account supports the sector-level trigger and impact. Revolut's own review remains the source for its architecture, recovery milestones and corrective actions.
For incident reporting, record the external trigger, pre-existing loss of redundancy, alerting failure, failed automatic failover, customer impact, manual recovery and remediation as separate facts. This prevents a supplier outage from becoming a complete explanation for why one firm recovered more slowly than another.
- Trigger: Google Cloud power loss in one London availability zone.
- Control weakness: reduced coordination-server redundancy was not alerted before the outage.
- Failure mode: the remaining servers could not form quorum, blocking automatic database failover.
- Recovery evidence: record service-by-service restoration times rather than one broad resolved time.
- Corrective action: test zonal failover under load and monitor quorum health, alert delivery and recovery tooling.
Worked practice example: an authorisation incident
This is a fictional training exercise, not a regulatory conclusion. At 09:10 a payments team sees a rise in failed card authorisations. Monitoring confirms that two UK customer groups are affected. The team does not yet know whether the cause sits with its gateway, a scheme connection or an external supplier.
At 09:20 the incident lead opens the evidence log. Operations records error rates and affected journeys. Customer support records reproducible symptoms without collecting full card details. Supplier management checks third-party status and contract contacts. Compliance maps the verified impact to the three thresholds and records the decision owner.
At 09:45 service begins to recover, but some transactions remain unresolved. The team updates the report with the recovery stage, customer actions and outstanding questions. It does not label the incident resolved until monitoring supports that status. A later root-cause review can correct the cause without rewriting the earlier timeline.
- 09:10: first observed symptom and monitoring source.
- 09:20: affected services, users, geography and transaction stage.
- 09:30: threshold assessment, uncertainty and accountable owner.
- 09:45: recovery evidence, customer action and next review time.
- Closure: final impact, confirmed cause, remediation and lessons.
Know what each payments role contributes
Operational reporting is a team activity. A candidate does not need to claim ownership of the final regulatory decision to show useful experience. The stronger example explains how their work improved the evidence or the response.
Operations staff can show monitoring, triage and handover discipline. Customer teams can show how they grouped symptoms and protected sensitive data. Product and engineering staff can explain dependencies and recovery evidence. Risk and compliance staff can show threshold assessment, governance and record keeping. Supplier managers can show escalation paths, service evidence and concentration-risk questions.
Build interview evidence without inventing an incident
Use a real example only if you can discuss it without exposing confidential customer, employer or security information. Explain the task, the evidence you checked, the decision boundary and the result. Remove names and sensitive technical detail where required.
If you do not have workplace experience, use the fictional exercise above. Label it as practice. Create a one-page incident log, a threshold assessment and a short handover. This demonstrates judgement without pretending that you made a regulatory filing.
- Separate observed facts from estimates and assumptions.
- Name the escalation trigger and the person authorised to decide.
- Show how you protected payment and personal data.
- Explain how you checked recovery and corrected the record.
- State what you would improve in the next exercise.
Questions to ask a prospective employer
Good questions reveal whether the organisation has turned policy into working practice. They also help you understand the role before accepting it.
- Who owns the decision that an incident meets a reporting threshold?
- How do operations, compliance, engineering and supplier teams share one timeline?
- Which services and third parties are mapped as critical dependencies?
- How are customer impact estimates checked and corrected?
- What exercises will the team complete before 18 March 2027?
Practise the distinction
| Skill | Payments task | Evidence |
|---|---|---|
| Triage | Confirm service, channel, geography and transaction stage | Timestamped incident log |
| Threshold assessment | Map verified impact to the three FCA thresholds | Reasoned assessment with an owner |
| Communication | State impact, status and next action without guessing cause | Clear internal or customer update |
| Supplier management | Check external dependencies and escalation routes | Dependency map and contact path |
| Recovery | Test stability and reconcile unresolved transactions | Closure checks and corrective actions |
| Dependency mapping | Compare provider-specific effects during one upstream outage | Component matrix with timestamps and evidence status |
Sources and dates
- FCA
Checked 31 Oct 2026, 13:48 UK time · Published 2026-03-18 - FCA Policy Statement PS26/2
Checked 31 Oct 2026, 13:48 UK time · Published 2026-03-18 - FCA Finalised Guidance FG26/3
Checked 31 Oct 2026, 13:48 UK time · Published 2026-03-18 - Bank of England PS7/26
Checked 31 Oct 2026, 13:48 UK time · Published 2026-03-18 - PXP Status
Checked 13 Nov 2026, 21:47 UK time · Published 2026-11-13 - Worldpay for Platforms Status
Checked 14 Nov 2026, 11:54 UK time · Published 2026-11-12 - Worldpay Status
Checked 14 Nov 2026, 11:54 UK time · Published 2026-11-12 - Associated Press
Checked 13 Nov 2026, 09:58 UK time · Published 2026-11-13 - Microsoft service health status
Checked 13 Nov 2026, 20:20 UK time · Published 2026-11-13 - ITV News
Checked 13 Nov 2026, 20:20 UK time · Published 2026-11-13 - Royal Bank of Scotland business service status
Checked 14 Nov 2026, 14:05 UK time · Published 2026-11-14 - Reuters
Checked 14 Nov 2026, 14:05 UK time · Published 2026-11-14 - Revolut
Checked 15 Nov 2026, 11:26 UK time · Published 2026-11-13 - Bank of England, PRA and FCA
Checked 15 Nov 2026, 11:26 UK time · Published 2026-11-04