CIS Controls Implementation Groups: Prioritizing with IG1, IG2 and IG3

How do you choose IG1, IG2 and IG3 in CIS Controls v8? A phased roadmap from the IG1 safeguards, measuring coverage, and the relationship with ISO 27001.
Technical summary
Technology
CIS Controls v8/v8.1, CIS CSAT, ISO/IEC 27001:2022, NIST CSF 2.0
Topic
Prioritizing a security program with CIS Controls implementation groups
Level
Beginner–intermediate
Prerequisites
  • A basic asset and account inventory
  • Access to reports from existing security tools
In this article7 sections

When pulling together a scattered set of controls, the real question is always the same: what gets done first? There’s a list hundreds of items long, an IT team of two and never-ending operational work. The CIS Critical Security Controls give an answer to this question that scales with the organization’s risk profile: Implementation Groups (IG). This guide is for IT managers and security leads and explains the difference between IG1, IG2 and IG3, choosing the right group, and how to turn the IG1 safeguards into a workable roadmap.

What CIS Controls are and what they’re for

CIS Controls is a prioritized set of cyber defense safeguards published by the Center for Internet Security. Version 8 drops the old device-centric structure and divides the safeguards into 18 activity-focused controls and the 153 safeguards beneath them. Version 8.1 keeps the control structure, clarifies definitions, aligns the safeguards’ security functions with NIST CSF 2.0 and adds the “Govern” function.

CIS Controls isn’t a management system standard. It doesn’t grant certification and doesn’t describe a risk assessment process. Its strength lies elsewhere: it gives a concrete answer, derived from common attack techniques, to the question “which technical safeguard, in what order?” It doesn’t conflict with ISO 27001; it’s used within it to sequence control implementation.

No. Check Focus
1–2 Inventory of enterprise assets and software What is being protected?
3 Data protection Data inventory, access, encryption
4 Secure configuration Baseline, default accounts
5–6 Account and access control management Account inventory, MFA, privileges
7 Continuous vulnerability management Scanning and patching
8 Audit log management Collection, retention, review
9–10 Email/browser protections, malware defenses Initial access vectors
11 Data recovery Backup and isolation
12–13 Network infrastructure management, network monitoring and defense Infrastructure and detection
14 Security awareness and skills training The human factor
15 Service provider management Supplier risk
16 Application software security The development process
17–18 Incident response management, penetration testing Response and validation

Implementation groups: IG1, IG2, IG3

Every security safeguard is assigned to one or more implementation groups. The groups are cumulative: IG2 includes all of IG1; IG3 includes all of IG1 and IG2. A higher group doesn’t skip the lower one.

  • IG1 – basic cyber hygiene: 56 safeguards. The minimum level for organizations with limited IT resources, mostly using off-the-shelf software, that need protection against opportunistic rather than targeted attacks. The safeguards were chosen so they can be implemented without specialist security knowledge, mostly by configuring existing products.
  • IG2: 74 safeguards on top of IG1, 130 in total. Organizations with multiple departments, systems with different risk profiles, and a separate IT team managing the infrastructure. Network monitoring, application security and more mature log management start in this group.
  • IG3: 23 additional safeguards, 153 in total. Organizations that process sensitive data, are subject to regulation and are likely to face targeted and sophisticated attacks. Safeguards requiring expertise, such as a penetration testing program, red teaming and advanced detection, are here.

The safeguards of some controls, such as network monitoring and defense and penetration testing, start at the IG2 or IG3 level. That doesn’t mean an organization targeting IG1 ignores these areas. The assumption is this: without basic hygiene, these safeguards don’t produce results. An IDS alarm on a network with no inventory is of no use to anyone.

Choosing the right group for the organization

The choice of group is determined less by the organization’s size than by the sensitivity of the data processed, regulatory obligations and the capacity of the security team. This sequence of questions works as a practical decision tree:

  1. Are the basic asset inventory, MFA, backup and patching processes documented and measurable? If not, the first stage is IG1, whatever the target.
  2. Do customer contracts, sector regulations or assessments such as TISAX require specific technical controls? If yes, the IG2 safeguards those requirements correspond to are added to the target.
  3. Does the organization have full-time security staff or a managed SOC service? If not, IG2’s monitoring-heavy safeguards can’t be sustained; even if the target is IG2, a service model should be planned for these safeguards. An alarm nobody looks at is a switched-off alarm.
  4. Are there systems whose outage or leak would affect public safety, critical production or the sensitive data of many people? If yes, IG3 safeguards are assessed for those systems.

A single group doesn’t have to apply to the whole organization. Setting IG2 as the general target and applying selected IG3 safeguards in an environment holding design data or in an internet-facing payment system is a legitimate approach. What’s required is that the reasoning for the choice is written down. It’s the first thing the auditor asks.

A roadmap from the IG1 security safeguards

Starting all 56 safeguards at once isn’t realistic; the team goes back to operations in the second week and the project stalls halfway. Planning four waves based on the dependencies between safeguards spreads the load and lets each wave produce data for the next.

Wave Safeguard area Example safeguards Output
1. Visibility Asset, software and account inventory 1.1, 2.1, 5.1 A scope list
2. Access MFA, administrator accounts, dormant accounts 5.3, 5.4, 6.3–6.5 The identity attack surface shrinks
3. Configuration and patching Secure configuration, automatic patching, malware 4.1, 7.3, 7.4, 10.1 Known vulnerabilities are closed
4. Resilience Backup, logs, incident response, awareness 8.2, 11.2, 11.4, 17.1 Recovery and response capacity

The visibility wave is the precondition for everything. A server not in the inventory isn’t in the patch report, the backup scope or the list of log sources either. A common picture: everyone says “we have an inventory”, but when you put the DHCP leases next to the EDR console, dozens of devices are missing from one of the two lists. In this wave, the goal isn’t a complete CMDB; it’s a list merged from discovery sources (DHCP, the EDR console, the directory, network scanning) with known owners.

In the access wave, the fastest risk reduction comes from MFA: make it mandatory for internet-facing applications, remote access and administrator access. For dormant accounts, CIS gives 45 days of inactivity as the reference. On the directory side, measuring it takes a single query:

Search-ADAccount -AccountInactive -TimeSpan 45.00:00:00 -UsersOnly |
  Where-Object { $_.Enabled } |
  Select-Object SamAccountName, LastLogonDate, DistinguishedName

The query relies on the lastLogonTimestamp attribute; this value replicates between domain controllers with a delay, so check borderline accounts once more before disabling them. In the configuration wave, a baseline is defined and applied; how to measure it at the device level is in the configuration compliance and drift tracking with CIS Benchmarks article. On the patching side, the safeguard isn’t “automatic patching is on”; it’s tracking the time to close. For prioritization and SLA design, see the vulnerability management program guide.

In the resilience wave, keeping at least one copy of backups isolated (11.4) and collecting audit logs (8.2) stand out. These safeguards don’t stop an attack on their own; they determine how quickly you recover after an incident and whether you can understand what happened. A backup server joined to the domain isn’t isolated.

Measuring the implementation level

A yes/no answer to “has the safeguard been implemented?” is misleading. A safeguard whose policy is written may not be working on half the systems. Assess each safeguard on two dimensions:

  • Policy/design: Is it defined how the safeguard will be implemented, and does it have an owner?
  • Implementation coverage: On how many of the in-scope assets does it really work? This value is measured from a tool: the EDR console, MDM compliance report, patch report, backup job report. Estimates aren’t written down.

For this assessment, CIS offers a tool called the CIS Controls Self Assessment Tool (CSAT); it records policy and implementation status per safeguard and tracks change over time. Even without the tool, a table with the same structure can be kept: safeguard number, owner, policy status, coverage (measured value and its source), date of last measurement, open work.

Mistakes in measurement we see again and again in the field:

  • Using the number of devices the tool itself sees as the denominator instead of the number of devices in the inventory. The EDR console only knows devices with the agent; that’s why “100 percent coverage” reports appear. Always compare the ratio against an independent inventory.
  • Treating a one-off measurement as permanent. Account, patching and backup safeguards change weekly, monthly at the latest.
  • Calculating ratios without noting the systems left out of scope (old servers, OT devices).

Relationship with other frameworks

CIS publishes official mapping documents between CIS Controls and ISO/IEC 27001:2022, NIST CSF 2.0 and NIST SP 800-53. Thanks to these mappings, CIS Controls can be used as the technical implementation list for a management system requirement.

Framework Nature Relationship with CIS Controls
ISO/IEC 27001:2022 A certifiable management system Prioritizes the technical implementation of Annex A controls
NIST CSF 2.0 An outcome-focused framework, profile approach Turns the subcategories in the target profile into concrete safeguards
CIS Benchmarks Product-specific configuration guides Used in implementing Control 4

In an organization using ISO 27001, the practical route is to write the relevant CIS safeguards next to each technical control in the Statement of Applicability (SoA). A general sentence such as “8.8 management of technical vulnerabilities is implemented” then breaks down into measurable safeguards. For the SoA structure, see the ISO 27001 SoA guide on the compliance side. On the CSF side, CIS safeguards form the technical evidence for the current and target profiles; an example is in the creating a NIST CSF 2.0 profile article.

The current text of the controls, implementation group assignments and mapping documents can be downloaded from the CIS Controls page. If you need an independent measurement of the current state and a prioritized work plan, a control and configuration audit covers the technical assessment, and enterprise cybersecurity consulting covers turning the roadmap into a program.

Frequently asked questions

Is IG1 for small companies, or is it meaningful for large organizations too?

IG1 is the minimum level for every organization. Even if a large organization targets IG2 or IG3, the IG1 safeguards are part of those groups; they must work with full coverage first. In large organizations, the problem usually isn’t the absence of a safeguard but that it doesn’t cover all assets.

What’s the difference between CIS Controls and CIS Benchmarks?

CIS Controls is a program that says which security safeguards to apply at the organizational level and in what order. CIS Benchmarks, on the other hand, are setting-by-setting secure configuration recommendations for a specific operating system, application or device; they’re used especially in implementing the secure configuration control.

Can you get certified with CIS Controls?

No, CIS Controls isn’t a certification scheme. The implementation level is measured by the organization’s own assessment or an independent assessment. If certification is needed, a management system standard such as ISO 27001 is used; CIS Controls supports its technical implementation.

When is it right to move up an implementation group?

A move to the higher group is planned when the safeguards in the current group work measurably across the full scope and the risk profile has changed, for example when a new regulation, sensitive data processing or a customer requirement comes up. Moving to a higher group while basic safeguards are missing spreads resources thin.

Support

Contact us for support or a discovery session on this topic.

Contact form
[email protected]

Eretek Cybersecurity

Firewall, VPN, endpoint security and threat prevention. We reduce the attack surface and measure and report the risks.

Knowledge Center

Other articles