- 1📋Living Risk RegisterUpdate on changes, incidents, and annual review; Git history as Metrics feed audit
- 2📊Operational MetricsTrack incidents, patch SLA compliance, and training completion cFindings go to leadership
- 3🔍Internal AuditAnnually test your own controls against the documented baseline Decisions drive change
- 4👔Management ReviewLeadership reviews metrics and findings, records decisions (clauChanges update the register
- 5🔄Improvement ActionsFeed decisions back into the ISMS to close the PDCA loop
The certificate is the starting line — a valid ISMS needs living evidence: audits, metrics, and reviews in a PDCA cycle.
🔎 Last updated · Editorial policy · Report an error
TL;DR: Passing the Stage 2 audit is the starting line, not the finish. An ISO 27001 certificate only stays valid if you can demonstrate a living ISMS — internal audits, management reviews, metrics that prove controls actually run, and a recertification audit every three years. This final part of the series covers how to operate the ISMS day-to-day, keep your risk register alive, and consolidate everything from Parts 1–5 into one checklist.
Authorization note: All commands below are meant to run against your own lab machines only — VMs you own or systems you have explicit written permission to audit. Never scan, probe, or pull logs from systems you don’t control.
What the Certificate Actually Commits You To
ISO 27001 certificates are typically issued for a three-year cycle. Within that cycle:
- Year 1: Certification (Stage 1 document review + Stage 2 implementation audit).
- Years 1 and 2: Surveillance audits — shorter on-site or remote audits, usually 6–12 months apart, checking that the ISMS still operates.
- Year 3: Recertification audit — a full re-assessment comparable in scope to the original Stage 2.
Surveillance audits focus on continuity: Are internal audits happening? Did management review occur? Are nonconformities from last time closed? Is the Statement of Applicability (SoA) current? If you’ve let the ISMS gather dust, this is where it shows.
Keeping the Risk Register Alive
Your risk register is not a document you finish — it’s a database you maintain. In practice, it should change whenever:
- A new system, vendor, or data flow is introduced (change management triggers a risk review).
- An incident occurs (post-incident review updates likelihood or impact scores).
- Annually at minimum, as part of management review — even if “no changes.”
Keep the register in something you can diff and query. A Git-tracked CSV works fine for a small ISMS:
~/isms/
├── risk-register.csv
├── soa.md
├── incidents/
├── internal-audits/
└── management-review/
$ git log --oneline risk-register.csv | head -5
a3f19c2 Q3 review: added vendor risk for new CRM
7b44e01 Post-incident update: R-014 likelihood raised to High
21ce88a Initial baseline risk assessment
The git history itself is audit evidence — it shows the register was touched throughout the year, not rewritten the night before the surveillance audit. Auditors appreciate that.
Metrics That Prove Operation
Surveillance auditors don’t want to see that you documented controls — they want evidence the controls ran. Collect metrics continuously; they feed both internal audit and management review.
Incident statistics
Track, per quarter: incident count, severity breakdown, mean time to detect (MTTD) and mean time to respond (MTTR). Even a quiet quarter is data — “0 incidents, monitoring verified by monthly alert test” is a valid record.
Patch SLA compliance
Define an SLA (e.g., critical vulnerabilities patched within 14 days, high within 30) and measure against it. On your own lab machines, a quick local check demonstrates the measurement method:
# List packages with pending security updates (Debian/Ubuntu)
$ apt list --upgradable 2>/dev/null | grep -i securi | wc -l
3
# Show kernel version vs. latest available
$ uname -r
5.15.0-91-generic
$ apt-cache policy linux-image-generic | grep Candidate
Candidate: 5.15.0-105.115
Record the output in a monthly patch-compliance file. The delta between runs is your metric. If criticals sit unpatched past SLA, either fix the process or document a risk-accepted exception — both are defensible; silence is not.
Training completion
ISO 27001 clause 7.2/7.3 requires competence and awareness. Track completion percentages per cycle. For a lab-scale ISMS, a simple CSV per quarter suffices:
quarter,person,training,completed,date
2025-Q3,alice,Phishing awareness,yes,2025-08-14
2025-Q3,alice,Access control refresher,yes,2025-09-02
2025-Q3,bob,Phishing awareness,no,-
The no row is important — auditors trust registers that show exceptions being chased, not 100% scores with no follow-up.
The Internal Audit and Management Review Loop
Plan-Do-Check-Act (PDCA) is the engine of continual improvement. Two Check activities are mandatory:
- Internal audit (clause 9.2) — at least once per certification cycle, but annually is the safe rhythm. Audit against your own scope and controls, document findings (conformities, nonconformities, observations, improvement opportunities).
- Management review (clause 9.3) — leadership reviews audit results, incident metrics, risk register changes, and improvement actions, then records decisions.
A lightweight internal audit you can run on a lab host — checking your own documented baseline against reality:
# Check: is SSH password authentication disabled per our baseline?
$ sudo sshd -T | grep -E 'passwordauthentication|permitrootlogin'
passwordauthentication no
permitrootlogin no
# Check: is the audit daemon running per our logging control?
$ systemctl is-active auditd
active
# Check: no world-writable files in a sensitive directory
$ find /srv/app -perm -o+w -type f | head
(no output — control conforms)
Save commands + outputs as the audit workpaper. One file per control checked. Findings feed the corrective action process (clause 10.2): root cause, correction, and a check that the fix worked.
Integrating with DPDP and GDPR Work
If you process personal data, ISO 27001 and privacy law overlap heavily. Don’t run two parallel programs — map them.
| ISO 27001 activity | DPDP (India) / GDPR overlap |
|---|---|
| Risk register entries for personal data flows | Records of processing / data mapping |
| Incident response process | Breach notification duties (GDPR: 72 hours to authority; DPDP: notify Data Protection Board and affected data principals) |
| Access control and logging controls | Security of processing (GDPR Art. 32) |
| Vendor/supplier controls in the SoA | Processor agreements, data processing addenda |
| Training program | Privacy awareness modules (one session can cover both) |
Practical tip: add a contains_personal_data column to your risk register. When the answer is yes, the same entry doubles as evidence for privacy compliance. One review, one meeting, two programs updated.
Surviving Surveillance and Recertification Audits
What auditors actually ask for at surveillance:
- The current SoA — has scope or control applicability changed?
- Internal audit report and evidence of closed findings.
- Management review minutes with decisions and actions.
- Samples: incident records, patch evidence, training logs, access reviews.
- Progress on any nonconformities from the previous audit.
A useful dry run is a quarterly self-check against the same list. Keep a single “audit binder” folder with dated evidence files; when the auditor asks, you produce it in minutes instead of reconstructing history for a week.
$ ls ~/isms/evidence/2025-Q3/
2025-09-30_access-review.csv
2025-09-28_patch-compliance.txt
2025-08-14_training-log.csv
2025-07-02_internal-audit-findings.md
2025-10-01_management-review-minutes.md
The Full From-Scratch Checklist (Parts 1–6)
The whole series in one table — your ISMS build plan:
| # | Step | Series part | Key deliverable |
|---|---|---|---|
| 1 | Define scope, context, and interested parties | Part 1 | Scope statement, boundaries |
| 2 | Leadership commitment and policy | Part 2 | ISMS policy, roles, objectives |
| 3 | Risk assessment and treatment | Part 3 | Risk register, treatment plan |
| 4 | Statement of Applicability and Annex A controls | Part 4 | SoA, control documentation |
| 5 | Operational controls, awareness, documentation | Part 5 | Runbooks, training records, asset inventory |
| 6 | Internal audit + management review | Part 6 | Audit report, review minutes |
| 7 | Stage 1 + Stage 2 certification audit | Part 6 (context) | Certificate, initial findings |
| 8 | Continual operation: metrics, surveillance prep | Part 6 | Quarterly evidence, KPIs |
| 9 | Yearly: risk register review, internal audit cycle | Part 6 | Updated register, audit reports |
| 10 | Year 3: recertification audit; restart cycle | Part 6 | Renewed certificate |
Defender takeaways
- Treat the register as a living database — track it in Git so change history doubles as audit evidence.
- Collect operation metrics continuously (incidents, patch SLA, training completion); quiet quarters still need documented verification.
- Never skip internal audit or management review — these are the first things surveillance auditors check, and they drive the corrective action loop.
- Map privacy work onto the ISMS — one risk register, incident process, and training program can serve DPDP/GDPR and ISO 27001 simultaneously.
- Maintain a dated evidence folder per quarter — producing surveillance evidence in minutes beats reconstructing it in weeks.
This closes the ISO 27001 from-scratch series. You now have the full path: scope, policy, risk, SoA, operations, and the post-certificate lifecycle. The certificate is a snapshot; the ISMS is the system. Keep the system running and the audits take care of themselves.
Related reading
ISO 27001 From Scratch
- What an ISMS Actually Is and How to Scope It
- Risk Assessment That Auditors Accept
- The Statement of Applicability Without Tears
- Evidence, Documents and the Trail Auditors Follow
- The Certification Audit — Stage 1, Stage 2 and What Fails
- Running the ISMS After the Certificate (Continual Improvement)You are here
Check your understanding
1. Why is a Git-tracked risk register valuable during a surveillance audit?
2. What do surveillance auditors primarily want to see regarding controls?
3. According to the article, how should a training record showing an incomplete item be handled?
4. If critical vulnerabilities remain unpatched past the SLA, which response is acceptable?
Your answers stay in this browser only.

