Part of ISO 27001: Critical success factors, a series on building a management system that reflects how your organisation really runs.
Bootstrapping a cycle
An ISO 27001 management system is a loop. The first time through, you're bootstrapping that loop: building just enough of each part to get it turning, then improving it on every pass.
A useful lens is Plan, Do, Check, Act. The current edition of the standard doesn't name it, but its structure follows it closely:
| Stage | Clauses | What it covers |
|---|---|---|
| Plan | 4 to 6 | Context, leadership, risk assessment and treatment, the Statement of Applicability |
| Do | 7 and 8 | Support, operating controls, running the risk process |
| Check | 9 | Monitoring and measurement, internal audit, management review |
| Act | 10 | Corrective action and continual improvement |
Before the loop starts there's one more step: Commit. Leadership agrees what they're signing up to, the scope is set and the forum is formed. Commit isn't a separate stage in the standard; it's a way of grouping the leadership and scoping work from clauses 4 and 5 that has to happen before anything else. It isn't repeated each cycle, but management review revisits it.
Don't rush Plan
The most common mistake is to hurry through Plan and jump into Do. It feels like progress, but it moves the cost downstream:
- controls designed without their evidence in mind turn every audit into an evidence hunt
- aspirational controls become nonconformities you wrote yourself, and each one needs corrective action
- controls with no risk behind them add more to monitor, measure and audit, for no benefit
- risks without real owners mean decisions get reopened at every forum
Front-load the work in design and the audits become far less stressful.
That doesn't mean staying in Plan for months. A good first Plan is thorough and small: a short risk register, the controls you already run, described well. Each later pass builds on it.
Who does what
| Stage | Who owns it | What that looks like |
|---|---|---|
| Commit | Leadership | Sponsor, scope, forum, top-level policy |
| Plan | A central team, with risk owners | Risk method, risk wording, control design, SoA, policies |
| Do | Existing operational teams | Running controls as part of normal work, producing evidence as they go |
| Check | The central team, with the forum | Metrics, internal audit, management review |
| Act | The forum decides; owners deliver | Corrective actions and an updated improvement plan |
The central team adds most value in Plan and Check. Do sits with the teams who already run the business, which is what "the business owns its controls" looks like in practice. Investing in Plan is what lets Do stay light: controls fit how teams already work, so running them is mostly normal work.
The stages
Commit
Brief your leadership on what they're signing up to and why. Agree a sponsor, set up the management forum, and agree the scope and top-level policy.
- Minimum: a named sponsor, a leadership briefing, terms of reference for the forum, an agreed scope and policy
- Ready to move on when: leadership has formally agreed the scope and the forum has met
- On later cycles: revisited through management review rather than repeated
See management commitment, scope and the management forum.
Plan
Agree how you'll assess risk, then sit down with managers, pen and paper, and ask what they're worried about. Describe the controls you already run, map them to Annex A, and write policies that match practice.
- Minimum: an agreed risk method; a first set of owned risks (ten to twenty is a good starting point); treatment decisions; controls described as what, when, where, who and why, with their evidence defined; a Statement of Applicability; topic policies
- Ready to move on when: the forum has reviewed the risks, owners have accepted residual risk, and the SoA and first improvement plan are approved
- On later cycles: targeted updates driven by events, metrics and audit findings, rather than a rebuild
See risk analysis, control selection and the SoA and policies and standards.
Do
Operational teams run the controls as part of normal work and file evidence as they go. Events are captured as they happen.
- Minimum: control owners who know their controls; evidence filed predictably; an event log in use
- Ready to move on when: there's enough operating history for the controls to be tested, which in practice usually means a few months of records
- On later cycles: new or changed controls from Act are absorbed into normal work
See control operation and security events.
Check
The central team gathers metrics and runs the internal audit; the forum carries out the management review.
- Minimum: a small set of agreed metrics; an internal audit covering clauses 4 to 10 and a sample of controls; a management review covering all the required inputs
- Ready to move on when: findings and decisions are recorded
- On later cycles: the audit programme rotates through the scope, and the forum reviews quarterly
See metrics, assurance and audit and the management forum.
Act
The forum turns what Check found into decisions, and owners turn decisions into actions.
- Minimum: findings, events and metrics converted into corrective actions in the improvement plan, with owners and dates
- Ready to move on when: the forum has approved the updated plan, and the next Plan pass starts from it
- On later cycles: this is the engine of continual improvement
See the risk improvement plan.
Choosing an iteration model
The right way to iterate varies by organisation. Two models cover most situations:
- Time-boxed. Each pass covers the whole scope at increasing depth, on a fixed cadence. It's a cleaner delivery model and suits a certification deadline.
- Area by area. Start with one part of the scope or the highest risks, then widen. It allows for better engagement: each area's leaders get real attention and their own decisions to talk about, and lessons from one area carry into the next.
A likely blend is a time-boxed first delivery that brings the whole scope up to a baseline, followed by cycles that work area by area, uplifting control definitions as they go.
When to go for certification
The key balance is how early to go for certification. Too early and you're auditing an ISMS that hasn't turned yet; too late and you're spending effort polishing rather than proving.
A good option is to engage a certification body early, and bring them in at the end of the first or second time box for a view on readiness before you schedule Stage 1. Many certification bodies offer a pre-assessment for this. They can tell you where you stand, but they can't advise you on how to fix it, because their accreditation requires them to stay impartial.
The certification audit itself comes in two stages:
- Stage 1 reviews your documentation and readiness. It can come once the first Plan is complete and Do is under way.
- Stage 2 tests whether the ISMS is implemented and working. It needs a full first cycle, including Check: the ISMS has to have run long enough to produce credible evidence, and certification bodies will normally expect that to include at least one internal audit and one management review.
Timing is paced by evidence rather than effort. Commit and the first Plan can take weeks. The cycle as a whole is paced by how long it takes to build up a few months of operating records.
Start on day one
One thing doesn't wait for the cycle: start capturing security events on day one. Even a simple log gives your first risk assessment real material to work with, and it's the earliest sign that the ISMS reflects how the organisation actually runs.
Further reading
- ISO 27001: Critical success factors: the main article, with the full diagram
- How to run an ISO 27001 ISMS on spreadsheets: a practical guide to the risk register, controls, evidence and SoA without a platform
Tom Whipp is the founder of RiskQuilt. He spent around thirty years in cyber and technology risk in UK financial services, including CISO and Director of Technology Risk roles.