- Scope & ContextClause 4: draw the ISMS boundary across activities, sites, asset
- Leadership CommitmentClause 5: signed policy, assigned roles, demonstrable resources
- Risk & SoAClause 6: risk register, treatment options, Statement of Applica
- Operate ControlsClauses 7–8: competence, awareness, run processes on a schedule
- Check & ImproveClauses 9–10: internal audit, management review, correct nonconf
An ISMS is a repeating PDCA loop of people, processes, controls, and records — documents are just evidence the loop ran.
🔎 Last updated · Editorial policy · Report an error
TL;DR: An ISMS is a management system — people, processes, and controls working together — not a folder of policy PDFs. In ISO 27001, your biggest early decision is the scope: what systems, locations, and people are inside the boundary, because that boundary determines exactly which evidence you must produce at audit time.
Housekeeping before we start: everything in this series is written for learners practicing on their own lab machines, their own cloud accounts, or written permission from the asset owner. Do not run discovery, scanning, or evidence-collection tooling against systems you don’t own. Part 1 is mostly paperwork, but later parts include asset-discovery commands — get authorization first.
Why This Series Exists
ISO 27001 has a reputation for producing binder-shelf certifications that do nothing for actual security. That happens when teams treat the standard as a writing assignment. It isn’t. ISO 27001 describes a management system: a repeatable loop where you decide what matters, manage risk, implement controls, check whether they work, and improve. The documents are just the fossil record of that loop.
This is Part 1 of a series. By the end you’ll understand what an ISMS actually is, how clauses 4–6 work in plain language, and how to write a scope statement that won’t haunt you at audit time.
What an ISMS Actually Is
An Information Security Management System (ISMS) is the sum of:
- People — named roles with security responsibilities (not “everyone is responsible,” which means nobody is).
- Processes — risk assessment, change management, incident response, access reviews, run on a schedule.
- Technology — the controls those processes operate: MFA, logging, encryption, backups.
- Records — evidence that the processes ran: review logs, risk registers, meeting minutes, ticket trails.
The classic failure mode is buying a policy template pack and calling it an ISMS. An auditor will ask two questions within the first hour: “Show me this policy operating” and “Show me evidence it ran recently.” A document pile answers neither.
The PDCA Cycle: The Engine of ISO 27001
ISO 27001 is built on Plan-Do-Check-Act, and the clause structure mirrors it:
- Plan — Clauses 4–6: understand context, define scope, lead the commitment, set objectives, assess risk.
- Do — Clause 7 (resources, competence, awareness) and Clause 8 (operate the processes, including risk treatment).
- Check — Clause 9: monitoring, measurement, internal audit, management review.
- Act — Clause 10: correct nonconformities and improve.
Annex A (mapped to ISO 27002) is the control catalogue — 93 controls in the 2022 version. But Annex A controls are selected based on your risk assessment; they’re not a fixed checklist you complete. Teams that start at Annex A instead of scope and risk end up implementing controls for risks they don’t have.
Clauses 4–6 in Plain Language
Clause 4: Context and Scope
Figure out your organization’s situation: what you do, who you serve, what laws apply, what interested parties (customers, regulators, insurers) expect. Then draw a line around what the ISMS covers. That’s the scope, and it must be documented and justified. Everything in this article hinges here.
Clause 5: Leadership
Top management must set the information security policy, assign roles, and demonstrably commit resources. “Demonstrably” means you can show it: a signed policy, a security sponsor on the org chart, budget decisions recorded. If the CEO has never heard of the ISMS, you don’t have one.
Clause 6: Planning — Risk, Objectives, Statement of Applicability
This is the analytical heart. You must:
- Establish a risk assessment methodology (how you score likelihood and impact).
- Run a risk assessment and produce a risk register.
- Choose risk treatment options (mitigate, accept, transfer, avoid) and map chosen controls into the Statement of Applicability (SoA) — the document that says which Annex A controls apply and why.
- Set measurable information security objectives (“patch criticals within 14 days,” not “be secure”).
Clauses 7–10 then operate, check, and improve what clauses 4–6 defined. Get 4–6 wrong and everything downstream inherits the error.
How to Define Scope — And Why It Decides Audit Difficulty
Scope has four dimensions you must state explicitly:
| Dimension | Question | Example |
|---|---|---|
| Business activities | Which processes? | SaaS platform delivery, customer support — not office catering |
| Locations | Which sites? | AWS eu-west-1, one London office — not the warehouse |
| Assets & systems | Which tech? | Production VPC, laptop fleet, identity provider |
| People | Who’s in the boundary? | 50 employees + 5 contractors with production access |
Shrinking scope shrinks the audit — legitimately or not. If you scope out the office network, no evidence is needed for it. But auditors check for scope gaming: if your out-of-scope warehouse network shares a domain controller with production, an auditor will extend the boundary via dependencies. The rule: you may exclude areas, but you must consider interfaces and dependencies on excluded areas and show the risk decision.
Practical guidance:
- Don’t over-scope to look impressive. Every included location/system adds evidence obligations (access reviews, asset inventories, physical controls).
- Don’t under-scope past your customers’ expectations. A customer handling your SaaS doesn’t care that “headquarters” is out of scope if the app lives entirely in-scope — but they’ll care if your billing system doesn’t.
- Shared infrastructure pulls things in. One AD domain, one IdP, one Jira used by both in-scope and out-of-scope teams? The auditor will follow the pipe.
Worked Example: Scoping a 50-Person SaaS Company
Meet Acme Metrics: 50 people, one office, a multi-tenant analytics SaaS on AWS (eu-west-1), corporate Google Workspace, and a part-time outsourced bookkeeper. Target: certification in 12 months.
Step 1: Inventory what exists
Before drawing the boundary, find out what’s actually there. In your own lab or cloud account, enumerate assets. For an AWS lab environment:
# List AWS regions with active resources (own account only)
aws ec2 describe-regions --output text --query 'Regions[].RegionName'
# Inventory running instances in your target region
aws ec2 describe-instances
--filters "Name=instance-state-name,Values=running"
--query 'Reservations[].Instances[].[InstanceId,InstanceType,Tags[?Key==`Name`].Value|[0]]'
--output table
Expected output:
----------------------------------------------------------------------
| DescribeInstances |
+---------------------+-----------+----------------------------------+
| i-0abc123def456 | t3.medium| acme-prod-app-1 |
| i-0def456abc123 | t3.small | acme-prod-worker-1 |
| i-0ghi789jkl012 | t3.micro | acme-staging-app-1 |
+---------------------+-----------+----------------------------------+
For on-prem lab machines, a quick host inventory:
# Linux: list listening services (what the ISMS boundary actually exposes)
ss -tlnp
# or on Windows PowerShell (own machine):
# Get-NetTCPConnection -State Listen | Select LocalAddress,LocalPort,OwningProcess
Step 2: Draw the boundary
Acme’s scoping conversation (recorded in a meeting minute — that’s evidence later):
- IN: AWS production + staging accounts, Google Workspace, all 50 laptops, the IdP (Google SSO), CI/CD pipeline, customer data flows, the London office network (because VPN access to prod terminates there — shared dependency).
- OUT: The bookkeeper’s own systems (interface considered: she receives CSV exports over a portal with MFA; that interface is in scope, her infrastructure isn’t), the sales CRM sandbox with no production data.
Note the bookkeeper decision isn’t just “excluded” — it’s excluded with a documented interface and risk decision. That’s the difference between scoping and scope gaming.
Step 3: Assign the roles
- Top management — the CEO signs the security policy, approves budget, and chairs the quarterly management review. Non-delegable.
- ISMS lead (often the security engineer at this size) — owns the risk register, SoA, documentation, and audit preparation. Must have authority, not just homework.
- Internal auditor — must be independent of the ISMS they audit. At a 50-person company, you can’t audit your own work, so options are: a person from another function trained in auditing, or a contracted internal auditor. Certification bodies accept both; self-auditing your own ISMS does not satisfy Clause 9.2.
- Everyone else — complete security awareness training and follow the processes. Clause 7.3 requires awareness evidence: keep completion records.
Scope Statement Template
Copy, adapt, and get management sign-off on this block:
ISMS SCOPE STATEMENT — Acme Metrics Ltd (template)
1. Organisation: Acme Metrics Ltd, 50 employees, London, UK.
2. Purpose of ISMS: Protect confidentiality, integrity, and availability of
customer analytics data and the services that process it.
3. Scope of the ISMS:
The development, operation, and support of the Acme Metrics
multi-tenant analytics SaaS platform, including:
Business activities:
- SaaS platform delivery and operations (AWS eu-west-1)
- Software development and CI/CD
- Customer support and account administration
Locations:
- AWS eu-west-1 (production and staging accounts)
- London office network and laptop fleet
Systems and assets:
- Production/staging VPCs and all workloads within them
- Google Workspace (identity provider, email, file storage)
- CI/CD tooling and source code repositories
- All corporate laptops (50 devices, centrally managed)
People:
- All 50 employees and contractors with access to in-scope systems
4. Exclusions and justification:
- Third-party bookkeeping infrastructure: OUT of scope. Interface
IN scope: encrypted CSV transfers via authenticated portal with MFA.
Risk decision recorded in risk register (RSK-014).
- CRM sandbox (no production data, no production connectivity):
OUT of scope; dependency review confirmed no shared identity or
data paths with in-scope systems.
5. Interfaces with excluded areas: reviewed and documented; see register
RSK-014 and the third-party management procedure.
6. Approval: signed by CEO (top management), dated, version-controlled.
Reviewed at each management review and on significant change.
Lab Exercise: Start Your Own ISMS Repository
Practice version-controlling your ISMS artifacts from day one — auditors love version history:
mkdir ~/isms && cd ~/isms
git init
mkdir 04-context 05-leadership 06-planning 07-support 08-operation 09-check 10-improvement
cp scope-statement-template.txt 04-context/scope-statement-v0.1.txt
git add .
git commit -m "ISMS init: clause structure + draft scope statement v0.1"
# Expected output:
# [main (root-commit) 3f2a1c9] ISMS init: clause structure + draft scope statement v0.1
# 8 files changed, 34 insertions(+)
Every future edit becomes a commit with a message — instant change history for Clause 8 and Clause 10 evidence.
Checklist
- Document your scope across all four dimensions (activities, locations, assets, people) with management sign-off — scope drives the entire audit surface.
- Document every exclusion with an interface and risk decision, not just a one-line “excluded.”
- Assign Clause 5 roles to named people: management sponsor, ISMS lead, and an independent internal auditor — never audit your own work.
- Inventory real assets before drawing the boundary so shared dependencies (one IdP, one domain) don’t drag surprises into scope mid-audit.
- Version-control ISMS artifacts from day one so you can prove the PDCA loop is actually turning.
In Part 2, we’ll build the risk assessment methodology and start filling the risk register.
Related reading
- Intercepting Desktop App Traffic with Burp Suite: System Proxy Setup, HTTPS Capture, and Beating Certificate Pinning
- Debug a .NET Thick Client with dnSpy: Decompile, Patch and Understand Obfuscated Desktop Apps
Check your understanding
1. Why does a purchased policy template pack fail as an ISMS?
2. How are Annex A controls supposed to be selected?
3. What happens if you scope out a system that shares dependencies with in-scope systems?
4. Why is defining scope called the biggest early decision?
Your answers stay in this browser only.



