Compliance · Risk Assessment

Practical Risk Assessment for Security and Compliance Programs

15 min read By Sushil Ojha
Risk AssessmentISO 27001NIST CSFGovernanceCompliance

Risk assessment is where security programs either become useful or become decorative. Done well, it forces uncomfortable clarity: what can hurt us, how badly, how likely, who owns the decision, and what we will do before the calendar moves on. Done poorly, it produces a colorful heatmap that nobody trusts and everyone updates the week before an audit.

Frameworks — ISO 27001, SOC 2, NIST CSF — expect a living method, not a ceremony. After enough assessments that aged into folklore, the lesson is simple: methodology beats eloquence, ownership beats adjectives, and residual risk without an expiry date is just deferred honesty.

What a useful risk assessment actually produces

Not a 90-page novel. A decision-grade register:

  • Scoped assets, processes, and threat scenarios that leadership recognizes
  • Consistent scoring for likelihood and impact — with written definitions people can argue about productively
  • Inherent risk, current controls, and residual risk stated without poetry
  • Treatment choices: mitigate, transfer, avoid, or accept — with named authority
  • Dates, owners, and next review triggers that survive reorganizations

Scope like an adult

Unbounded scope creates unbounded theater. Start with business services and data classes that matter: customer-facing systems, identity, payment or clinical data paths, crown-jewel intellectual property, and critical vendors. Include people and process failure modes, not only CVE catalogs. Exclude explicitly — and document why — so auditors do not assume you forgot the basement servers.

A practical method that scales

  1. Context — Business objectives, regulatory obligations, risk appetite, and in-scope environments
  2. Identify — Threats, vulnerabilities, and scenarios tied to assets and processes
  3. Analyze — Likelihood × impact using a published scale; note control effectiveness honestly
  4. Evaluate — Compare residual risk to appetite; escalate what exceeds it
  5. Treat — Action plans with owners, due dates, and success criteria
  6. Monitor — Trigger-based and calendar-based reviews; update after incidents and material change

Keep scoring simple enough that two competent analysts land near the same result. If every assessment requires a philosopher, the scale is broken.

Likelihood and impact without self-deception

Likelihood should reflect exposure, threat activity, control strength, and historical incident data — not optimism. Impact should reflect confidentiality, integrity, availability, regulatory exposure, and customer trust. A “medium” that means different things in different departments is how risk registers become fiction.

Write the scale down. Train people on examples. Revisit when the business changes. Do not let “critical” become a political badge.

Treatment decisions that hold up

  • Mitigate — Control projects with measurable completion, not vague “improve monitoring”
  • Accept — Named approver, rationale, compensating controls, expiry, and revisit date
  • Transfer — Insurance or contractual shift that actually transfers something
  • Avoid — Stop doing the risky thing; rare, powerful, underused

Eternal risk acceptance is not acceptance — it is abandonment with stationery.

Cadence that prevents museum registers

  • Full assessment — At least annually, or on major architecture/service change
  • Trigger reviews — After incidents, new regulated data flows, material vendor changes, or control failures
  • Register review — Quarterly with security and business stakeholders who can fund treatment

Minimum evidence set

  • Scoped risk methodology document people can find without archaeology
  • Risk register with owners, dates, treatment status, and residual scores
  • Statement of risk appetite and acceptance authority
  • Traceability from high risks to remediation projects or operating controls
  • Meeting notes or approvals showing management review — not only security’s private spreadsheet

Common failure modes

  • Copy-pasting last year’s register with new dates
  • Scoring only technical vulns while ignoring process and third-party scenarios
  • No link between “high residual risk” and budget or backlog
  • Heatmaps that cannot explain a single decision under cross-examination

Closing thought

A good risk assessment is a management instrument. It should change priorities, funding, and control design. If it only changes a slide deck, you performed documentation — not risk management. Build a method you can defend, a register you can operate, and a review cadence that keeps residual risk honest.

← Work & Write-ups