Skip to content
DT Nova

Access control policy: roles, permissions matrix, joiner-mover-leaver and a template

An access control policy sets out who may use which systems and data, on what basis access is granted, how it changes when people join, move or leave, and how it is reviewed. A good access control policy follows least privilege, groups permissions into roles, and is enforced by the systems themselves rather than by goodwill. This guide covers the principles, role-based and attribute-based access control, an example permission matrix, the joiner-mover-leaver process, reviews, the legal requirements and a template outline.

Sources were read on 12 November 2026: Directive (EU) 2022/2555 (NIS2), Regulation (EU) 2016/679 (GDPR), NIST's role-based access control project page, NIST SP 800-162 on attribute-based access control, and the OWASP Top 10:2025.

Why access control matters

Many security incidents start with access someone already had: an employee's account taken over by phishing, a former contractor's login that was never removed, a shared admin password, a service account with far more rights than it needs. In applications, access control is the most common weakness: in the OWASP Top 10:2025, broken access control holds the first place, and OWASP reports that 100% of the applications tested were found to have some form of it. A clear policy, applied in every system, limits what any single compromised account can do.

The principles

  • Least privilege. Each person and system gets only the access needed for their tasks, for as long as they need it.
  • Need to know. Access to confidential data follows the task, not seniority or curiosity.
  • Separation of duties. The person who creates a supplier does not approve its payments; the person who writes code does not approve its own release to production.
  • Roles, not individuals. Permissions are given to roles; people are assigned to roles.
  • Personal accounts. No shared logins; every action can be traced to a person or a named service account.
  • Strong authentication. Multi-factor authentication for all remote and privileged access, and for every system that holds personal or business-critical data.
  • Default deny. Access not explicitly granted is refused, in systems and in network rules.

Role-based and attribute-based access control

Role-based access control (RBAC) is the standard approach in business systems. NIST's RBAC model, developed with Sandhu, Ferraiolo and Kuhn, was adopted as American National Standard ANSI/INCITS 359-2004 on 11 February 2004 and revised as INCITS 359-2012. Roles mirror jobs: "sales representative", "sales manager", "accounts payable", "warehouse supervisor". When a person moves from sales to purchasing, their role changes and their permissions follow, without anyone hunting for individual rights.

Attribute-based access control (ABAC) decides access from attributes: of the user (department, location, clearance), of the resource (owner, classification, customer) and of the environment (time, device, network). NIST SP 800-162, published in January 2014 and updated in 2019, provides a definition of ABAC and considerations for using it. ABAC handles rules like "a sales representative sees only the customers in their own region" or "patient records are visible only to the care team of that patient". Most systems use both: roles for broad permissions, attributes for row-level rules.

Example: a permission matrix for a mid-sized company

RoleCRMERP ordersERP financeHR systemCloud admin
Sales representativeOwn region: editCreate, view ownNoneOwn dataNone
Sales managerAll regions: editView all, approve discountsReports onlyOwn team: viewNone
Accounts payableNoneViewEnter invoicesOwn dataNone
Finance managerReports onlyViewApprove paymentsOwn team: viewNone
HRNoneNonePayroll exportAll: editNone
IT administratorConfigurationConfigurationConfigurationConfigurationNamed admin account, MFA, logged

The matrix is the heart of the policy: one table per system or one table across systems, owned by the business, with every role, every permission and the reason. Notice that the IT administrator configures systems but does not approve payments or read HR data in normal work; separation of duties applies to IT as well.

Joiner, mover, leaver

  1. Joiner. HR records the start date and job; the role follows from the job; accounts are created from the role before day one; multi-factor authentication is set up at first login.
  2. Mover. When the job changes, the old role is removed and the new one added on the same day. Movers who keep old rights "for a transition" are the usual source of excessive access.
  3. Leaver. Accounts are disabled at the end of the last working day, or immediately on a dismissal; devices returned; shared secrets the person knew are changed; the account deleted after the retention period.
  4. Contractors and suppliers. Access with an end date by default, renewed only on request.

The process works only if HR events reach IT automatically. Connecting the HR system to the identity system so that roles follow job data is one of the most valuable integrations a company can build.

Privileged access

Administrator accounts deserve their own rules: separate admin accounts from everyday accounts, multi-factor authentication always, as few people as possible, access granted for a limited time when needed, every session logged, and break-glass emergency accounts stored securely and checked after every use. Service accounts for integrations get exactly the rights the integration needs and credentials stored in a secrets manager, never in code.

Reviews and logging

Rights drift over time. Regular reviews bring them back: the owner of each system receives a list of who has which role and confirms or removes each entry. Privileged access is reviewed more often than ordinary access. Logs of logins, permission changes and access to sensitive records are kept for an agreed period under the data retention policy, and reviewed when something looks wrong.

What the law requires

NIS2 Article 21(2)(i) lists "human resources security, access control policies and asset management" among the minimum risk-management measures for essential and important entities, and Article 21(2)(j) adds the use of multi-factor or continuous authentication where appropriate. See our NIS2 compliance guide.

Under the GDPR, Article 32(4) requires the controller and processor to take steps to ensure that any person acting under their authority who has access to personal data does not process it except on instructions from the controller. Article 25(2) requires measures ensuring that, by default, personal data are not made accessible without the individual's intervention to an indefinite number of people. Role-based permissions, row-level rules and access logs are how software meets both. The access control policy is one of the topic policies under the information security policy.

An access control policy template outline

  1. Purpose, scope, owner, approval, version, review date.
  2. Principles: least privilege, need to know, separation of duties, default deny.
  3. Roles and the permission matrix per system, with data owners.
  4. Authentication: password rules, multi-factor authentication, single sign-on.
  5. Joiner, mover and leaver procedures, with time limits.
  6. Privileged and service accounts.
  7. Remote access and access by suppliers.
  8. Access reviews: frequency, owners, recording.
  9. Logging and monitoring of access.
  10. Exceptions and how they are approved and expire.

Access control in your own software

In the systems we build, access control is designed in from the start: roles that match your organisation, row-level rules where data must be separated by team, region or customer, single sign-on with your identity provider and enforced multi-factor login, permission changes and sensitive reads logged, and automated tests that check each role sees exactly what it should. That is part of every web app and internal tool we deliver, hosted in the EU, with the code in your repository. A security audit checks the same in systems you already run.

Questions people ask

What is an access control policy?

An access control policy states who may access which information and systems, on what basis, how access is granted, changed and removed, and how it is checked. It turns principles such as least privilege and separation of duties into rules that IT and system owners apply every day.

What is role-based access control?

Role-based access control (RBAC) gives permissions to roles, such as sales, accounting or warehouse supervisor, and assigns people to roles, instead of giving permissions to each person individually. When someone changes jobs, they change role and their permissions follow. NIST's RBAC model was adopted as an American national standard in 2004.

What is the difference between RBAC and ABAC?

RBAC decides access by the user's role. Attribute-based access control (ABAC) decides by attributes of the user, the resource and the context, for example department, the customer's country, the record's classification or the time of day. Many systems combine them: roles for the broad permissions, attributes for finer rules.

How often should access rights be reviewed?

Privileged and administrator access at least every quarter, other access at least twice a year, and always when someone changes role or leaves. The review is done by the system or data owner, who confirms or removes each right, and the result is recorded.

Is an access control policy required?

NIS2 lists access control policies among the minimum cybersecurity risk-management measures for essential and important entities. The GDPR requires appropriate security and steps to ensure that people with access to personal data process it only on instructions. In practice, auditors, customers and insurers expect one from any company handling sensitive data.

Sources