To document ISO 27001 controls properly, you need three things: the documented information the standard explicitly requires in clauses 4 to 10, a Statement of Applicability that links every Annex A control to a decision and to real evidence, and records that prove each applicable control actually operates. Policies alone are not enough, and a separate procedure for each of the 93 controls is rarely necessary.

Documenting controls is where most implementations waste time or get the focus wrong. The common problem is not a lack of documentation; it is documentation that does not reflect what actually happens in the organisation, or that came from generic templates nobody on the team recognises as their own.

A certification auditor does not value the thickness of the binders. They value the coherence between what the policies say, what the records show and what the team answers when asked. This guide explains how to reach that coherence efficiently: what is mandatory, how to structure the Statement of Applicability, how to document each type of control, what auditors look for, and how to keep documentation alive after certification.

What ISO 27001 means by "documented information"

ISO/IEC 27001:2022 uses the term documented information instead of the classic "documents and records". It covers any information the organisation must control and maintain, in any format or medium: a policy in a wiki, a ticket in a service desk, a log exported from a cloud console or a signed PDF. Understanding what ISO 27001 contains, clause by clause, makes it much easier to see which documented information is required and which is your choice.

Documents and records

Although the standard uses a single term, documented information serves two different functions:

  • Documents describe how things are done: policies, procedures, methodologies and standards. They are written once and updated when a process changes.
  • Records prove that what the document says was actually done: minutes, logs, tickets, review evidence and training completions. They are generated continuously in day-to-day operations.

The most frequent mistake is creating many documents and few records. An auditor can live with a brief policy if the records demonstrate the process works. They cannot pass a system full of perfect policies with no evidence that anyone follows them.

What clause 7.5 requires for every document

Clause 7.5 sets out how documented information must be created and controlled. In practice, every document in your ISMS should have:

  • Identification: a title, owner, version and date, so anyone can tell which version is current.
  • Review and approval: evidence that someone with authority reviewed it and confirmed it is suitable.
  • Availability: the people who need it can find it when they need it.
  • Protection: access is restricted where appropriate, and the document cannot be changed without control.
  • Version control: changes are tracked, and obsolete versions are withdrawn or clearly marked.
  • Retention rules: records are kept for a defined period and disposed of securely.

Documents from outside the organisation that the ISMS relies on, such as supplier security reports, contracts or regulatory texts, must also be identified and controlled.

How much documentation is enough

The standard says the extent of documented information depends on the size of the organisation, the complexity of its processes and the competence of its people. A 40-person SaaS company and a 4,000-person bank both need a scope, a policy, a risk assessment and a Statement of Applicability, but the length and number of supporting procedures should be very different.

For the Annex A catalogue of 93 controls, the requirement is not to write a procedure for every control. It is to be able to demonstrate that each applicable control exists and works. Often a short policy plus records from the tools you already use is the most convincing evidence you can give.

Why the Statement of Applicability (SoA) is the ISMS's central document

The Statement of Applicability is the document that connects the Annex A controls to your organisation's reality. Clause 6.1.3 d makes it mandatory, and it is usually the first document a certification auditor asks to see, because it shows in one place which controls you claim to operate and where the proof is.

What each row of the SoA should contain

For each of the 93 Annex A controls, a useful SoA records:

  • The control reference and name from the 2022 version.
  • Whether the control applies, and the justification for including or excluding it.
  • Whether it is implemented, and to what degree.
  • The risk, legal requirement or contractual obligation that the control addresses.
  • The policy, procedure or technical evidence that supports it, with a precise location.
  • The control owner and the date of the last review.

A well-built SoA is not a spreadsheet with 93 rows of "applies / implemented" and nothing else. Each row must be traceable to something real: a procedure, a system or a record. If you cannot make that connection for a control marked as implemented, you have a potential nonconformity before the audit begins.

Linking the SoA to risk treatment

The SoA and the risk treatment plan must tell the same story. Every control selected to treat a risk should appear in the SoA as applicable, and every applicable control should be traceable to at least one risk, legal requirement or contractual obligation. If a control is marked as applicable but no risk or requirement explains why, the auditor will ask what it is protecting.

Auditors often test this link directly. They pick a high risk from the register, follow it to the treatment decision, find the controls chosen to reduce it, check those controls in the SoA and then ask for the evidence. Any break in that chain, such as a control missing from the SoA or an SoA row with no evidence, becomes a finding.

The simplest way to keep both documents aligned is to give each risk and each control a unique reference and use those references in both places. When the risk assessment is updated, the SoA is reviewed in the same exercise rather than months later.

Justifying exclusions

Do not exclude controls for convenience. An exclusion needs a solid reason, usually that the risk does not exist in your scope or that the activity is not performed at all. "Not relevant" or "not applicable" on its own is not a justification, and auditors review exclusions with particular attention because they are where scope problems usually hide.

Good justifications are specific to your context. A company that does not develop software may justify excluding secure coding (8.28) and outsourced development (8.30). A fully remote company may exclude some controls on physical perimeters if it has no offices in scope, but it still needs controls for devices used off-premises (7.9), storage media (7.10) and clear screens (7.7), because the risks to information on laptops do not disappear.

Be careful with exclusions that depend on suppliers. Hosting everything with a cloud provider does not remove physical security from your responsibility; it changes how you evidence it, typically through the provider's certifications and your own supplier review rather than through your own access logs.

Use precise cross-references

Point to evidence exactly. If control 5.9 (inventory of information and other associated assets) is covered by your asset management tool, name the tool and say where the records can be seen. A reference such as "see asset inventory, IT folder" forces the auditor to search; a reference to a specific register, report or dashboard lets them verify in seconds.

Where flexible workspaces are in scope, a desk booking system can support desk handovers and clear-desk routines under control 7.7, but it does not replace the control evidence itself. The same applies to any tool: it can generate or store evidence, but the SoA should still state which policy defines the control and which records prove it operates.

Keep the SoA alive

The SoA must be updated when the scope changes, when a new system or supplier is introduced, when a relevant process changes or when the risk assessment is revised. It is a living record of decisions, not a document produced once for the first audit.

Keep a version history and record who approved each version and why it changed. An SoA dated a year before the audit, with no changes despite a growing business, new tools and new suppliers, raises immediate questions about whether the ISMS is really being managed.

What documentation is mandatory by clause

Beyond the SoA, the standard requires documented information at specific points in clauses 4 to 10. This is the minimum no ISMS can go into a certification audit without:

Clause

Mandatory documented information

Practical tip

4.3 Context

ISMS scope

Describe boundaries, interfaces and dependencies, not just a product name.

5.2 Leadership

Information security policy

Keep it short, approved by top management and available to staff and relevant interested parties.

6.1.2 Planning

Risk assessment process and criteria

Define likelihood, impact and acceptance criteria before running the assessment.

6.1.3 Planning

Risk treatment process and Statement of Applicability

Make every treatment decision traceable to the SoA.

6.2 Planning

Information security objectives

Make them measurable and assign an owner and a deadline.

7.2 Support

Evidence of competence

Keep training records, certifications and role descriptions for security roles.

8.1 Operation

Information needed to show processes were carried out as planned

Usually covered by operational records such as tickets and logs.

8.2 and 8.3 Operation

Risk assessment and risk treatment results

Keep dated versions to show the risks are reviewed over time.

9.1 Evaluation

Monitoring and measurement results

Report a small set of KPIs consistently rather than many inconsistently.

9.2 Evaluation

Internal audit programme and results

Show the auditor's independence and the follow-up of findings.

9.3 Evaluation

Management review results

Minute the decisions and actions, not just the attendance.

10.2 Improvement

Nonconformities and corrective actions

Record the root cause and verify the action was effective.

Two of these documents tend to cause most trouble for first-time implementers. The risk methodology must be repeatable and produce comparable results, which is why many organisations start with a simple 5×5 risk matrix before adding sophistication. And the evidence of competence must cover all staff with security responsibilities, which is easiest to manage through a structured ISO 27001 employee training plan with dated records.

Everything else, including operating procedures, technical instructions, incident records and access logs, is needed to evidence controls, but the standard does not prescribe a specific format. That is good news: you can adapt the level of detail to your organisation.

How to document each Annex A control without drowning in paperwork

The catalogue of 93 Annex A controls can seem overwhelming. The key is understanding that different types of control need different types of documentation. Grouping controls by how they operate lets you choose the lightest documentation that still convinces an auditor.

Process controls

Controls such as access rights (5.18), incident management (5.24 to 5.28) and change management (8.32) are processes that people run repeatedly. They need:

  • A procedure describing who does what, when and with what authorisation.
  • Records proving the process runs: access request and removal tickets, incident logs with timelines, change tickets with approvals.

For example, documenting access management usually means a short access control policy, a joiner-mover-leaver procedure and the tickets or identity provider logs that show accounts were created, changed and removed as described.

Technical controls

Controls such as cryptography (8.24), technical vulnerability management (8.8), backup (8.13) and logging (8.15) are mostly implemented in systems. They need:

  • A documented technical decision: which algorithms, what frequency, what scope and which systems.
  • Technical evidence that it is running: configuration exports, automated reports, scan results and restore tests.

Vulnerability management is a good illustration. The policy defines scanning frequency and remediation deadlines by severity, and the evidence is the scan history plus tickets showing that critical findings were fixed on time. A clear vulnerability mitigation process makes that evidence consistent from one audit to the next.

Organisational controls

Controls such as supplier relationships (5.19 to 5.22), cloud services (5.23) and periodic reviews need:

  • A procedure establishing the frequency, criteria and owner.
  • The record that it was done: review minutes, contracts with security clauses, supplier assessments.

Supplier controls are where many first audits find gaps, because contracts and reviews are spread across procurement, legal and IT. A single vendor compliance management process with one register of critical suppliers keeps that evidence in one place.

People and physical controls

People controls such as screening (6.1), terms and conditions of employment (6.2), awareness and training (6.3), the disciplinary process (6.4) and confidentiality agreements (6.6) are usually evidenced through HR records. These include signed contracts and NDAs with security clauses, background check confirmations where legally allowed, onboarding checklists and training completions.

The documentation for these controls is often already in place, but held by HR rather than by the ISMS owner. A short note in the controls matrix saying which HR system holds which records, who can access them and how long they are retained is usually enough, without copying personal data into the ISMS.

Physical controls (7.1 to 7.14) are evidenced through office access logs, visitor records, clear desk checks and secure disposal certificates. For infrastructure hosted in the cloud, they are evidenced through your provider's certifications, combined with your own controls for the offices and devices that remain your responsibility.

Build a controls matrix

An efficient way to organise all of this is a controls matrix: a table where each row is an Annex A control and the columns capture what covers it and where the evidence lives. Well maintained, it serves both as the ISMS reference and as the internal audit guide.

Control

Covered by

Evidence

Frequency

Owner

5.18 Access rights

Access control policy, joiner-mover-leaver procedure

Identity provider logs, quarterly access review minutes

Continuous and quarterly

IT lead

8.13 Information backup

Backup standard

Backup job reports, restore test records

Daily and quarterly test

Infrastructure lead

5.22 Supplier service review

Supplier management procedure

Annual supplier review records, certificates on file

Annual

Operations lead

Control selection and documentation must be aligned with the ISO 27001 risk assessment: every control selected in the risk treatment plan must appear in the SoA and in the matrix with its evidence.

What the auditor really looks for in control documentation

Understanding the auditor's perspective saves many hours of unnecessary work. An ISO 27001 certification auditor is not looking for documentary perfection; they are looking for evidence of a working system.

5 questions for every control

For each control declared as implemented in the SoA, auditors effectively work through the same five questions:

  1. Does the policy exist? Is there a document establishing the organisation's intention for this control?
  2. Is it approved and current? Does it have an approval date, an owner and a next review date?
  3. Do people know about it? Is there evidence that relevant staff have received it or been trained on it?
  4. Is it applied in practice? Do records demonstrate the control was applied during the audited period?
  5. Is there measurement or periodic review? Does someone check that the control is working correctly?

If you can answer yes to all five for every implemented control, you are in good shape for the audit. A useful exercise before Stage 1 is to run through the questions control by control with each owner, using the controls matrix as the checklist, and note any answer that relies on memory rather than a record.

How auditors test documentation

Auditors rarely read every document. They sample. A typical approach is to choose a period and ask for specific records from it: the access review from a given quarter, the change tickets from a given month, the onboarding records of two recent joiners or the evidence for an incident mentioned in the management review. They then interview the people who run the control and compare what they say with what the documents and records show.

That is why records must be dated, findable and complete, not reconstructed the week before the audit. The logic is the same across security frameworks, and anyone who has had to prepare for a NIS2 audit will recognise the focus on coherence between paper and practice.

The most common finding

In a first certification, the most common finding is not the absence of a technical control. It is the absence of evidence for a control that does exist but that nobody recorded: an access review done verbally in a meeting, a restore test with no record, a supplier assessment kept in someone's inbox.

These findings are frustrating because the work was done, but from the auditor's point of view an unrecorded control cannot be verified. Moving from "we do it" to proving it with evidence is entirely preventable with a minimal, well-designed recording system from the start: a template, a fixed location and an owner for every recurring activity.

How PrivaLex helps you document ISO 27001 controls

PrivaLex supports compliance teams, CISOs and management in implementing ISO/IEC 27001 with documentation that is useful from day one: not binders of policies nobody applies, but documented controls that hold up in conversation with the auditor and that the team recognises as its own.

Documentation gap assessment

We start by reviewing what you already have, from policies and wikis to tickets, logs, contracts and HR records, and compare it with the mandatory documented information in clauses 4 to 10 and the Annex A controls in scope. We also interview the people who run the main processes, because the gap between what is written and what is done is usually where audit risk sits.

Many organisations have more evidence than they think; the gap is usually structure, ownership and dates. You receive a prioritised list of what needs to be written from scratch, what needs to be adapted and what simply needs to be referenced from existing tools, so effort goes where it changes the audit outcome.

SoA and controls matrix design

We build the Statement of Applicability and the controls matrix together with your team. Every inclusion and exclusion gets a justification tied to your risks and scope, every applicable control gets a precise evidence reference, and every control gets a named owner who understands what they are responsible for.

The result is a single map that the team, the internal auditor and the certification auditor can all follow, and that stays aligned with the risk treatment plan as the business changes. It also becomes the basis for reusing documentation if NIS2, ENS or SOC 2 are added later.

Documents that fit how you work

We draft policies and procedures around your real processes, tools and team size, and keep them as short as the auditor will accept. A ten-person engineering team and a regulated group with several business units need very different documents, even when the controls are the same.

Where your tools already generate evidence, such as an identity provider, a ticketing system, a code repository or an HR platform, we design the documentation to point to those records instead of creating parallel registers. Control owners review every document before it is approved, so the team recognises it as its own and can explain it in an audit interview.

Evidence and audit readiness

We set up minimal templates for recurring records, such as access reviews, supplier reviews and restore tests, and organise an evidence pack by clause and control so audit requests can be answered quickly. We then run the internal audit with the independence the standard requires and support leadership through the management review.

Before Stage 1 and Stage 2, we check that every row of the SoA can be followed through to dated evidence, and we rehearse auditor interviews with control owners so they can explain what they do and show the proof with confidence.

Maintenance and other frameworks

After certification, we support the annual review cycle, surveillance audits and scope changes, and help keep the calendar of recurring evidence on track. When the organisation needs to cover more than one standard, we extend the same documentation to ENS, NIS2, ISO 27701 or DORA through a shared control map, so each control is documented once.

For startups and scaleups, the same approach is described in our guide to ISO 27001 certification for startups, and the Kloutit case study shows how an AI company certified ISO 27001 with two people from its team involved and without slowing product development.

Formats that work (and those that do not)

The format of your documentation has a direct impact on whether it is used and maintained. These patterns come up again and again in audits.

Formats that work

Formats that work share two qualities: they are easy for the team to keep up to date, and they make evidence easy for an auditor to find.

  • Policy and procedure in one short document. For simple operational controls, a two-page policy that includes the basic procedure is more maintainable than a twenty-page document with sections nobody reads.
  • Records in existing tools. If access management lives in the corporate directory, the directory's logs are your access control records. You do not need an additional register.
  • Minimal templates for repeatable processes. For the quarterly access review or the annual supplier review, a template with the minimum fields ensures evidence is always captured, without depending on the memory of whoever runs it.
  • One document per topic, not per control. A single access control policy can cover several related controls (5.15 to 5.18, 8.2 and 8.5), with the SoA pointing each control to the relevant section.

As a rule of thumb, if a document cannot be reviewed by its owner in under an hour, it is probably too long for a small or mid-sized organisation. Short documents get read, updated and followed; long ones get approved once and forgotten.

Formats that do not work

The formats that fail in audits usually fail for the same reason: they describe an organisation that does not exist.

  • Documents downloaded from the internet without adaptation. Auditors recognise generic templates, and the team does not recognise them as its own, so it does not apply them.
  • Policies without an owner or review date. A policy without a last review date is treated as out of date by default.
  • Procedures that describe an ideal process nobody follows. If the procedure says "the security manager reviews access monthly" and the records show a single review in two years, that is a major nonconformity.
  • Evidence scattered across personal folders and inboxes. If records cannot be found within minutes during the audit, they effectively do not exist.

When a procedure and reality diverge, it is usually better to change the procedure to match a realistic process than to promise a frequency or level of detail the team cannot sustain. An honest quarterly review that always happens is worth more than a monthly review that rarely does.

Where to keep your documentation

Many organisations certify with well-structured shared drives or wikis such as Google Drive, SharePoint or Confluence, using folder permissions and version history to meet the control requirements of clause 7.5. This works well when the ISMS is small, the folder structure mirrors the clauses and controls, and one person is responsible for keeping it tidy.

Others use a GRC or compliance automation platform that links controls, evidence and reminders, and collects some technical evidence automatically. The right choice depends on the size of the ISMS and how much evidence can be gathered from systems; comparing AI compliance tools helps separate what software genuinely saves from what still needs human judgement.

Whatever you choose, keep one source of truth for each document, restrict who can edit approved versions and avoid copies in personal folders or email attachments, which quickly become outdated versions in circulation.

Why maintenance is the real challenge of ISO 27001 documentation

Many organisations reach certification with reasonably good documentation and lose it at the first surveillance audit because nobody updated it. Maintenance is the part of the ISMS most frequently underestimated, and it is exactly what surveillance and recertification audits are designed to test.

Formalised annual review

Once a year, before the surveillance audit, the owner of each policy and procedure checks whether it is still accurate. A "reviewed by / date" field and a short record of what changed, or a note that nothing needed to change, are enough to show the review happened.

Plan the review a few weeks before the audit rather than the week itself, so any updates can be approved and communicated in time. Combining it with a structured ISO 27001 readiness checklist helps make sure no area of the ISMS is skipped.

Updates on relevant changes

Every time a system, supplier or process with a security impact changes, the owner of the affected control updates the documentation. New tools, migrations, reorganisations and new suppliers are the most common triggers, and each can make a procedure or an SoA reference inaccurate overnight.

The most reliable way to make this happen is to add a "documentation impact" question to the change management process, so updates are triggered by the change itself rather than by memory. The same question should prompt a check of the risk assessment, because a significant change can introduce new risks as well as new documents.

Nonconformity and corrective action register

Every finding from an internal audit, external audit, incident or detected deviation must be recorded with its root cause, the corrective action taken and a check that the action worked. Closing a finding by fixing the symptom, without understanding why it happened, usually means it reappears at the next audit.

This is not only a clause 10 requirement; it is the primary evidence that the ISMS improves over time, which is what justifies recertification. A register with a handful of well-handled findings is more convincing to an auditor than one that is suspiciously empty.

A calendar of recurring evidence

Most records are generated on a schedule: quarterly access reviews, annual supplier reviews, backup restore tests, vulnerability scans, training refreshers, internal audits and management reviews. Put them on a shared ISMS calendar with owners and reminders, and check monthly that nothing has slipped.

A missed quarterly review discovered in month two can be recovered; one discovered by the auditor cannot. The same discipline applies to other regulatory programmes, and the principles for staying compliant with NIS2 transfer directly to ISO 27001.

Reusing ISO 27001 documentation across other frameworks

If your organisation works with more than one framework, ISO 27001 control documentation can be reused efficiently. Designed well, one policy and one set of records can satisfy several standards at once.

Where documentation overlaps

Most security and privacy frameworks ask for similar things in different words: a risk-based approach, documented policies, access control, incident handling, supplier oversight and evidence of review. That overlap is what makes reuse possible.

  • NIS2: the risk-management measures in Article 21, such as incident handling, business continuity, supply chain security, access control and cryptography, map closely to Annex A controls, although NIS2 adds specific incident reporting timelines and management accountability.
  • ENS: Spain's National Security Framework (Royal Decree 311/2022) has its own catalogue of measures in Annex II and its own Statement of Applicability, but much of the ISO 27001 documentation can be reused; an ENS certification project usually builds on it.
  • DORA: financial entities must document an ICT risk management framework and review it at least once a year, which can be built on ISO 27001 risk and control documentation.
  • GDPR: the ISMS risk analysis and security measures support Article 32, and can feed data protection impact assessments under Article 35 when processing involves personal data.
  • SOC 2: many controls and much of the evidence overlap, although the report format and the criteria are different.

The differences matter as much as the overlaps. Each framework has requirements that ISO 27001 does not cover, such as NIS2 and DORA incident notification deadlines or ENS security categories, so reuse should start from a mapping exercise rather than an assumption that ISO 27001 documentation is enough.

Design a multi-framework control map

The key is designing from the start with a control map: a layer that connects each Annex A control to the equivalent requirements of the other frameworks in scope. Each control is documented and evidenced once, and the map shows which framework requirements it satisfies. This avoids duplicated work, keeps documentation consistent and reduces the nonconformity surface when several auditors arrive in the same period.

When US customers ask for SOC 2 as well, the comparison of ISO 27001 vs SOC 2 shows which controls are shared and where the relevant differences lie.

Conclusion

Documenting ISO 27001 controls is not a writing exercise. It is an exercise in coherence between what the organisation decides, what the team does and what gets recorded. The organisations that pass audits smoothly are rarely the ones with the most documents; they are the ones where every applicable control can be followed from the risk it treats, through the Statement of Applicability, to a dated record produced by normal work.

The practical approach is consistent across organisations of any size. Start with the mandatory documented information from clauses 4 to 10. Build the SoA and a controls matrix that point each control to precise evidence. Keep documents short and specific to how your teams work, and let your existing tools generate the records. Then design maintenance from the start, with owners, review dates and a calendar of recurring evidence, so the documentation survives surveillance and recertification audits rather than being rebuilt each year.

The earlier that coherence is designed in, the less rework is needed before the audit, and the easier it becomes to extend the same documentation to NIS2, ENS, DORA or SOC 2.

Frequently Asked Questions (FAQs)

The standard does not set a fixed number of documents. It requires documented information at specific points in clauses 4 to 10: scope, policy, risk methodology, SoA, internal audit results and management review, among others. For Annex A controls, the requirement is to be able to demonstrate they exist and work, not necessarily with a specific procedure for each one.

Not every control needs to apply. Those that are excluded must be justified in the SoA. Those that do apply must be backed by some form of evidence, not necessarily a dedicated procedure; it can be a technical record, a documented configuration or a reference to an existing tool.

The SoA records for each Annex A control whether it applies, whether it is implemented and what evidence supports it. It must be formally approved by management or the ISMS owner with sufficient authority. It is one of the first documents the auditor requests in the opening hours of a certification audit.

At minimum, a formalised annual review of all policies and procedures. Additionally, an immediate update whenever the process, system or supplier that the control covers changes. Records, evidence that the control is working, are generated continuously in operations; they are not periodically updated documents but traces that accumulate over time.

To a large extent, yes. ISO 27001 Annex A controls overlap significantly with many NIS2 and ENS requirements. With a well-designed multi-framework control map, ISO 27001 documentation and evidence also cover equivalent requirements from other frameworks, avoiding duplicated work. The main differences are usually in specific notification requirements and sector-specific controls.

The most common is the gap between the procedure and real practice: the document states something is done at a certain frequency or in a certain way, but the records show it did not happen that way. Second is the absence of records for controls that do work technically but that nobody documented. Both are preventable if the recording system is designed from the start as part of the process, not as an extra step.

Free checklist
Do you know what’s standing between you and ISO 27001 certification?
Download our readiness checklist and find out which controls you already have in place and where your real gaps lie, before you start the process.
Download Free Checklist