Vendor Risk Management Framework: How to Manage Third-Party Risk with ISO 31000, COSO and ISO 27001
Summary:
- Third parties now feature in nearly half of all breaches
- VRM follows the ERM cycle, using contracts and evidence as its main tools
- Vendor risk tiering comes from data access, system links, service dependency and ease of replacement
- Critical vendors need proof of controls alongside questionnaires
- Linked CMDB, IAM, procurement and incident data enables continuous vendor monitoring
- Five critical vendors and four questions make a practical starting point
Ask a company how many vendors it has and you will usually get a confident number. Ask which five of them could stop the business for a week, and the room goes quiet. Vendor risk management (VRM) exists to close that silence before an incident does it for you.
Introduction
On the night of Friday 19 September 2025, a ransomware attack hit Collins Aerospace’s MUSE platform, the software several airlines use to run check-in, baggage drop and boarding at shared airport desks. Heathrow, Brussels and Berlin were the best-known airports affected, with Dublin and Cork also reported. Staff went back to manual check-in. Brussels asked airlines to cancel 140 of its 276 scheduled departures for the Monday because the vendor could not yet deliver a secure new version of the system. The EU cybersecurity agency ENISA confirmed ransomware was behind it. None of the airports had been breached directly. They had all bought the same service from the same supplier.
It is not an isolated pattern. Verizon’s 2026 Data Breach Investigations Report found a third party involved in 48% of breaches, up from 30% the year before and roughly 15% the year before that. The same report found that only 23% of third parties fully fixed missing or badly configured multi-factor authentication on their cloud accounts. The perimeter most organizations defend is no longer the one that matters. A large part of it belongs to someone else. This guide sets out a practical vendor risk management framework, built on ISO 31000 and COSO, that you can run as part of your wider governance, risk and compliance (GRC) program.
What Is Vendor Risk Management?
Vendor risk management is the practice of knowing which external parties your business relies on, how much each one could affect you, and what you are doing to keep that risk in check. It applies to suppliers, cloud and software providers, outsourced service teams, consultants and any partner with access to your data, systems or customers.
You will often hear it called third-party risk management (TPRM), and a third-party risk management framework is simply the structure that organizes it. Some organizations use TPRM for a wider group that also includes partners and fourth parties, meaning the vendors your vendors depend on.
How Vendor Risk Management Fits into Enterprise Risk Management and GRC
It is not a separate discipline from ERM. It is ERM applied to the part of the business you do not control. In GRC terms, it is the third-party layer of governance, risk and compliance. ISO 31000’s process and COSO’s components still apply: set appetite, identify, assess, respond, monitor and report. What changes is that your levers are contracts, evidence and oversight instead of direct control. Several standards then add the detail:
- ISO/IEC 27001:2022 has five Annex A controls on suppliers: information security in supplier relationships (A.5.19), supplier agreements (A.5.20), the ICT supply chain (A.5.21), monitoring and change management of supplier services (A.5.22) and cloud services (A.5.23). The ISO/IEC 27036 series goes deeper on supplier relationships.
- NIST CSF 2.0 added a Govern function in 2024, with cybersecurity supply chain risk management as one of its categories.
- DORA, the EU’s Digital Operational Resilience Act, applies to financial entities from 17 January 2025. It requires them to keep a register of their arrangements with ICT third-party providers and to manage those providers’ risk formally.
The common thread is that responsibility stays with you. A contract can share the cost of an incident. It does not share the accountability to your customers, your regulator or your board.
The Vendor Risk Management Process
The vendor risk management process fits neatly into the ERM cycle. The table lines up each stage with ISO 31000, COSO and the vendor-specific controls.
| Stage | ISO 31000 and COSO ERM | Vendor risk practice |
|---|---|---|
| Set the frame | ISO: scope, context and criteria (6.3). COSO: risk appetite and objectives (Principles 6–9) | Third-party risk policy, appetite, tiering criteria, named owners |
| Identify | ISO: risk identification (6.4.2). COSO: identifies risk (Principle 10) | Vendor inventory, dependency mapping to business services, fourth parties |
| Assess | ISO: analysis and evaluation (6.4.3–6.4.4). COSO: assesses severity, prioritizes (Principles 11–12) | Due diligence by tier, evidence review, inherent versus residual risk |
| Treat | ISO: risk treatment (6.5). COSO: risk responses, portfolio view (Principles 13–14) | Contract terms (ISO 27001 A.5.20), remediation plans, exit and fallback plans |
| Monitor | ISO: monitoring and review (6.6). COSO: Review and Revision (Principles 15–17) | Ongoing monitoring of supplier services (ISO 27001 A.5.22), reassessment on change |
| Report | ISO: recording and reporting (6.7). COSO: Information, Communication and Reporting (Principles 18–20) | Portfolio view, board reporting, regulator registers such as DORA’s |
Set the frame and set up vendor tiering
Start with a policy and an appetite: what kinds of vendor risk the organization will accept, and who can accept them. Then set up vendor risk tiering. A common basis is how much data the vendor holds, how deeply it connects into your systems, how much of a business service depends on it, and how hard it would be to replace. Spend decides nothing here. A vendor paid very little can still run a process your revenue depends on, as the airport check-in case shows. Tiering is the same step as ISO 31000’s criteria and COSO’s appetite, applied to suppliers, and it decides how much effort every later step deserves.
Identify the dependencies, including the hidden ones
A vendor list is not a dependency map. The useful question is not “who do we pay?” but “which of our business services stop if this vendor stops?” That is where your asset and service data matters (more below). Look one level further as well. Fourth-party and concentration risk, many of your vendors relying on the same cloud, the same software component or the same platform, is the kind of exposure no single assessment will reveal.
Assess in proportion: vendor risk assessment by tier
ISO 31000 and COSO both ask for risk to be analysed and ranked against appetite. For vendors, that means vendor due diligence matched to the tier: a short questionnaire for low-risk suppliers, evidence review, certifications and sometimes an on-site or remote audit for critical ones. A questionnaire only records what a vendor says about itself, so the vendor risk assessment should include evidence where it counts: audit reports, test results, penetration-test summaries, proof of controls such as MFA. Record both inherent risk, the exposure before controls, and residual risk after them.
Treat: contract, remediate, plan the exit
ISO 31000’s treatment options map onto vendor decisions more directly than most teams realize.
- Change the likelihood: require remediation of specific findings before go-live, or within a set time, with security obligations written into the agreement (ISO 27001 A.5.20).
- Change the consequence: insist on tested fallback procedures, data backups you control, and a realistic recovery time commitment. The airports’ manual check-in worked because a manual fallback existed, but it ran at a fraction of normal capacity.
- Share the risk: liability terms, audit rights, breach-notification deadlines and insurance.
- Avoid or exit: decline the vendor, or leave one. An exit plan that has never been exercised is a hope, not a plan.
- Retain by informed decision: accept the residual risk and record who accepted it. COSO’s portfolio view (Principle 14) matters here, because ten small accepted risks can add up to one large one.
Vendor Risk Monitoring and Strategic Alignment
A vendor passes due diligence once, at onboarding, when everyone is paying attention. After that the vendor changes its software, its staff, its hosting and sometimes its owners. The 2026 DBIR adds a reminder of how slowly findings get closed: the typical vendor fixed leftover cloud-account weaknesses far too slowly for comfort, and most third parties never fully resolved their MFA gaps. This is why continuous vendor monitoring matters as much as the first assessment.
What the frameworks ask for
ISO 31000 asks for monitoring and review through every stage of the process (6.6), and ISO 27001 turns that into a specific control for supplier services, covering regular monitoring, review and change management (A.5.22). COSO’s Review and Revision component adds a trigger test: assess whether substantial change has altered the risk picture (Principle 15), review performance (16) and keep improving (17). For vendors, substantial change is easy to name: a merger, an outage, a new subcontractor, a new data type, a contract renewal, a breach at the vendor’s other customers.
Indicators worth tracking
As with enterprise risk, a few indicators beat a dashboard nobody reads. Tie each one to the appetite you set at the start, with an owner and a threshold:
- Share of critical vendors with a current, in-date assessment.
- Open remediation items past their due date, by vendor tier.
- Critical vendors without a tested exit or fallback plan.
- Concentration: how many critical services rely on the same provider.
- Vendors with privileged access to your systems where MFA is not confirmed.
- Incidents and SLA breaches attributed to vendors, per quarter.
Integration: where the data comes from
Vendor risk gets stale fastest when the data is typed in by hand. The facts you need already live in other systems, and connecting them turns an annual exercise into continuous vendor monitoring.
- CMDB: which configuration items, applications and business services each vendor supports. This is what answers “what breaks if vendor X fails?” and makes tiering objective.
- Procurement, ERP and the contract repository: the vendor master record, spend, owner, contract terms, renewal and notice dates.
- Identity and access management: which vendor accounts exist, what access they hold, and whether MFA is enforced. This links straight to the DBIR finding on third-party authentication gaps.
- ITSM and incident tools: vendor-related incidents, changes and SLA performance, to compare what vendors promise with what they deliver.
- Security ratings and threat intelligence feeds: external signals about a vendor’s exposure between formal assessments.
- Finance and credit data: early warning on a vendor’s financial health.
- ISMS, audit and compliance tools: findings, control tests and evidence, so vendor oversight sits next to your own control assurance.
Two cautions. First, the vendor master record needs one owner and one source of truth. If procurement, security and legal each keep their own list, integration will simply copy the disagreement. Second, a CMDB that doesn’t record vendor-supplied services, or is out of date, will give you a clean-looking map that is wrong.
Report the tip, keep the iceberg
Boards do not need five hundred vendor scores. They need the critical vendors, how exposure is trending, what sits outside appetite, which concentrations worry the team, and what decisions are being asked of them. A unified GRC platform helps here: keep the per-vendor evidence underneath, one click away. Where a regulator requires a register, as DORA does for financial entities, the same data should feed both.
A Practical Example
Go back to the airports, and treat it as a vendor-risk case rather than a news story. Many airports used a common-use platform from one supplier. When the supplier was hit, the problem arrived at every customer at once, and each customer’s own security was beside the point. The fallback was pen and paper, and the recovery depended on the vendor’s ability to rebuild and re-issue a secure version, which the vendor could not do quickly enough in Brussels.
Four questions any vendor risk management program should be able to answer before an event like that:
- Which business services depend on this vendor, and which systems carry them? (The CMDB link.)
- What is the maximum time each of those services can be down before the loss goes beyond tolerance?
- Is there a fallback, has it been exercised, and at what capacity does it run?
- Who else among our vendors relies on the same provider, and what happens then?
If the answers take weeks to assemble, that is the finding. Organizations that run vendor risk this way tend to see the same benefits:
- Fewer surprises, because critical vendors are known, tiered and watched.
- Faster, calmer incident response, because fallback plans and contacts already exist.
- Stronger contract and audit positions, because requirements were set up front.
- Less duplicated work between procurement, security, legal and risk.
Conclusion
Vendor risk is enterprise risk with a different set of levers. ISO 31000 gives the process, COSO gives the strategic frame, and ISO 27001, ISO 27036, NIST CSF 2.0 and regulation such as DORA fill in the supplier-specific detail. Integrated data keeps the whole framework current.
Three things to take away:
- Tier vendors by the business services that depend on them, not by spend.
- Test fallback and exit plans for critical vendors, and monitor them against thresholds linked to appetite.
- Feed the program from your CMDB, procurement, identity and incident systems, with one owner for the vendor record.
If you want a starting point, list your five most critical vendors and ask: what service would stop, how long can it be down, what is the fallback, and when did anyone last test it? The gaps in those answers are your first project.
Need help connecting vendor, contract, procurement and identity data in one place? Talk to the Xponential Digital team.