SP 800-53B tells you which of the 1,000-plus SP 800-53 controls and enhancements to start from, and most tailoring mistakes come from treating that starting point as the finish line. The baselines (Low, Moderate, High and a separate Privacy baseline) are the basis for federal system authorisations, FedRAMP, and many state and commercial programmes that borrow them. This guide explains how a system lands on a baseline, what is actually in each one, the six tailoring activities NIST defines, how overlays and organisation-defined parameters fit, and the identity controls that the baselines leave out. It is for security architects, ISSOs, assessors and cloud providers who must select and justify a control set, and it assumes you know the SP 800-53 catalogue itself.
Key takeaways
- The impact level picks the baseline. FIPS 199 categorisation and the FIPS 200 high-water mark decide Low, Moderate or High; one High objective makes the whole system High.
- The baselines grow fast: 149 controls and enhancements at Low, 287 at Moderate and 370 at High in the Release 5.2.0 OSCAL profiles, most of the growth being enhancements.
- Tailoring is six defined activities, not deleting what you dislike. Every change must be risk-based, documented in the security plan and defensible to the authorising official.
- Baselines rest on stated assumptions. NIST says they do not address insider threat, advanced persistent threats or classified information; if those apply, you must supplement.
- Some identity controls are in no baseline. IA-9 Service Identification and Authentication is unallocated, so machine-to-machine authentication must be added deliberately.
What SP 800-53B is and its current status
NIST SP 800-53B, Control Baselines for Information Systems and Organizations, was split out of SP 800-53 in Revision 5 so that the catalogue of controls and the selection of controls could evolve separately. It was published in September 2020 and updated on 10 December 2020. On 27 August 2025 NIST issued Release 5.2.0, a minor release that, according to NIST, made no changes to the baselines and aligned the version number with SP 800-53 and SP 800-53A. The machine-readable baselines are published as OSCAL profiles in NIST's oscal-content repository; the figures on this page come from the Release 5.2.0 profiles.
The publication provides three security baselines, one per system impact level, and one privacy baseline that applies irrespective of impact level. It also gives tailoring guidance, the assumptions behind the baselines, and the method for building overlays. It sits in step 3 (Select) of the Risk Management Framework: see the NIST Risk Management Framework (SP 800-37) Guide. For the catalogue and its 20 families, see the NIST SP 800-53 Guide.
From categorisation to baseline
Categorisation under FIPS 199 assigns Low, Moderate or High potential impact to confidentiality, integrity and availability for the information types on the system. FIPS 200 then applies the high-water mark:
- Low-impact system: all three objectives are Low.
- Moderate-impact system: at least one objective is Moderate and none is High.
- High-impact system: at least one objective is High.
Controls in SP 800-53B are not allocated by security objective, so a system that is High for availability only still starts from the High baseline. Tailoring can then downgrade controls that support only one objective (NIST's example is CP-2(1) Coordinate with Related Plans, which supports availability). FedRAMP is stricter: its baselines page states that FedRAMP baselines do not allow tailoring of controls by confidentiality, integrity or availability. See the NIST SP 800-53 Privacy Controls and FIPS 199/200 Categorisation Guide for categorisation in detail.
What is in each baseline
The table counts controls and enhancements in each Release 5.2.0 OSCAL baseline profile, with selected families to show where the growth sits.
| Baseline | Total | Base controls | Enhancements | AC | IA | CP | SR |
|---|---|---|---|---|---|---|---|
| Low | 149 | 131 | 18 | 11 | 16 | 6 | 11 |
| Moderate | 287 | 177 | 110 | 39 | 24 | 23 | 12 |
| High | 370 | 188 | 182 | 46 | 26 | 35 | 14 |
| Privacy | 96 | 75 | 21 | Privacy-focused controls including PT and PM family controls; independent of impact level | |||
Three things stand out. First, the jump from Low to Moderate is mostly enhancements (18 to 110), so a Moderate system is not "Low plus a few controls". Second, contingency planning (CP) roughly sextuples from Low to High, which is why availability-driven High categorisations are expensive. Third, no Program Management (PM) controls appear in the three security baselines; PM controls are implemented at organisation level, and several appear in the Privacy baseline instead (for example PM-9 Risk Management Strategy).
The Privacy baseline is built differently. Of its 96 entries, 24 are PM controls and 13 are PT (PII Processing and Transparency) controls, with the remainder drawn from families such as IR, SI, SA and CA. SP 800-53B notes that it addresses privacy programme responsibilities under OMB Circular A-130, that it is not itself mandated by law or by A-130, and that implementing it does not by itself satisfy every OMB obligation. Privacy enhancements also require their base control to be selected, so a privacy team cannot adopt an enhancement while the security baseline has tailored out the parent control.
Identity controls by baseline
| Control | Low | Moderate | High | Why it matters |
|---|---|---|---|---|
| IA-2(1) and IA-2(2) MFA to privileged and non-privileged accounts | Yes | Yes | Yes | MFA for organisational users is required at every level |
| IA-2(8) Access to Accounts: Replay Resistant | Yes | Yes | Yes | Rules out replayable one-time codes for account access |
| IA-2(12) Acceptance of PIV Credentials | Yes | Yes | Yes | Federal PIV acceptance |
| IA-5(1) Password-based Authentication | Yes | Yes | Yes | Password rules apply wherever passwords remain |
| IA-5(2) Public Key-based Authentication | No | Yes | Yes | Certificate validation and private key protection, including for machine certificates |
| IA-5(6) Protection of Authenticators | No | Yes | Yes | Covers stored secrets and keys |
| IA-3 Device Identification and Authentication | No | Yes | Yes | Devices authenticate before connecting |
| IA-12 Identity Proofing | No | Yes | Yes | Proofing before credential issue |
| AC-2(2) Automated Temporary and Emergency Account Management, AC-2(3) Disable Accounts | No | Yes | Yes | Removes temporary and inactive accounts, including forgotten service accounts |
| AC-2(12) Account Monitoring for Atypical Usage | No | No | Yes | Behavioural monitoring of accounts |
| AC-6(9) Log Use of Privileged Functions, AC-6(10) | No | Yes | Yes | Privileged activity logging and enforcement |
| IA-9 Service Identification and Authentication | No | No | No | Not in any baseline: add it where services authenticate to each other |
The six tailoring activities
SP 800-53B names six tailoring activities. Each one produces a documented decision in the system security and privacy plan (PL-2), and PL-11 Baseline Tailoring is itself a control in the catalogue.
| Activity | What it means | Example | Evidence to keep |
|---|---|---|---|
| Identifying and designating common controls | Controls provided by the organisation or a platform and inherited by the system | Physical controls (PE) inherited from a data centre or cloud provider | Common control provider, inheritance list, provider's authorisation |
| Applying scoping considerations | Deciding where a control applies, based on placement, environment, technology, mission, security objective, legal and policy considerations | Wireless controls scoped out of a system with no wireless; CP-2(1) downgraded on a High-for-confidentiality-only system | Scoping rationale per control |
| Selecting compensating controls | Alternative controls that give equivalent protection when a baseline control cannot be implemented | A jump host with session recording where a legacy system cannot support MFA directly | Rationale, residual risk accepted by the authorising official |
| Assigning organisation-defined parameter values | Filling in assignment and selection operations | Inactivity period in AC-2(3), audit retention | Parameter values in the plan; NIST says they become a permanent part of the control |
| Supplementing baselines | Adding controls or enhancements the baseline does not include | Adding IA-9 for service-to-service authentication; adding AC-2(12) at Moderate | Threat or requirement driving each addition |
| Providing specification information | Extra detail needed to implement a control | Naming the approved MFA methods for IA-2(1) | Implementation statements |
Compensating controls: the rules
- Source: select from the SP 800-53 catalogue first; go outside it only if nothing suitable exists.
- Rationale: explain how the compensating control gives equivalent protection.
- Risk: the authorising official assesses and accepts the residual risk.
- Limits: NIST states compensating controls are not used to avoid the need to comply with requirements, and more than one may be needed.
Assumptions, gaps and overlays
The baselines rest on assumptions listed in SP 800-53B section 2.3: information is relatively persistent; systems are multi-user, networked and general purpose; some information is not shareable among authorised users; and the organisation has the structure and resources to implement the controls. NIST also lists what the baselines do not address, including insider threat, advanced persistent threats, classified information, information needing specialised protection and cross-domain communication. If any of these apply, supplementation is expected rather than optional.
Overlays package tailoring for a community, technology or environment so that each system does not repeat the same analysis. NIST's Security and Privacy Control Overlay Repository (SCOR) collects overlays in government-wide, public and NIST categories. National security systems use the CNSSI 1253 baselines rather than SP 800-53B. See the NIST Control Overlays Guide for overlay content and stacking, and the NIST SP 800-234 HPC Security Overlay Guide for a worked example.
Choosing an approach
| Situation | Recommended approach | Trade-off |
|---|---|---|
| US federal system | SP 800-53B baseline from FIPS 199, tailored and documented, plus required overlays | Tailoring needs authorising official agreement |
| Cloud service selling to US agencies | FedRAMP baseline (Low, Moderate, High or LI-SaaS), which builds on the Rev 5 baselines with FedRAMP parameters | No tailoring by security objective; FedRAMP sets many parameter values |
| Commercial organisation adopting 800-53 voluntarily | Moderate baseline as a reference, tailored by risk, with common controls designated at platform level | Freedom to tailor, but no external authority to accept risk, so governance must |
| Contractor handling CUI | SP 800-171 Rev 3, which is derived from the Moderate baseline, rather than 800-53B directly | Fewer requirements, different assessment regime |
Small versus large organisations: a small team should invest most in common controls (inherit as much as possible from a cloud provider's authorised platform) and a short, well-justified tailoring record. A large agency or provider with many systems should publish an organisational overlay that sets parameter values and standard supplements once, then tailor per system only for real differences. See the FedRAMP Guide and the NIST SP 800-171 and CMMC 2.0 Guide.
Implementing a baseline step by step
- Categorise: identify information types, assign impact per objective, apply the high-water mark and get the categorisation approved.
- Load the baseline as data: import the OSCAL profile for the chosen baseline into your GRC tool or a git repository rather than copying a spreadsheet, so later NIST releases can be compared.
- Designate common controls: list what the organisation and any underlying platform provide, with their authorisation status.
- Apply organisational overlays: load agency, sector or technology overlays, and the organisation's own parameter values.
- Scope: record each scoping decision with its consideration type (technology, placement, security objective and so on).
- Supplement: add controls for threats outside the baseline assumptions; review IA-9, AC-2(12) and SR controls for systems with heavy machine-to-machine traffic.
- Document: write the tailored set, parameter values and rationale into the system security and privacy plan.
- Review with the authorising official before build: agree compensating controls and accepted residual risk early, not at assessment.
- Revisit: re-run tailoring when categorisation, architecture or threat changes, and track tailoring decisions under configuration control so you can roll back a decision that proves wrong.
Where identity fits
Most baseline identity controls were written with people in mind, and systems pass assessment while their service accounts, API keys and workload identities sit outside the evidence. Treat the following as standard supplements or specification information for any system with automation:
- IA-9 Service Identification and Authentication: add it wherever services, pipelines or agents call each other; specify mTLS or workload identity tokens. See the mTLS Guide.
- AC-2 for non-human accounts: specify that account management, owners and AC-2(3) disabling apply to service accounts, with the inactivity parameter set for them too.
- IA-5(2) and IA-5(6): state how machine certificates and stored secrets are protected and rotated.
- AC-6(9): include privileged use by automation accounts in privileged function logging.
See the Machine Identity, PKI and Certificate Lifecycle Guide and the NHI Ownership Guide.
Detection and measurement
- Tailoring rationale coverage: percentage of tailored-out or compensated controls with a recorded rationale and accepted risk (target 100%).
- Parameter completeness: percentage of organisation-defined parameters with an assigned value in the plan.
- Inheritance validity: percentage of inherited common controls whose provider authorisation is current.
- Supplement drift: number of systems missing the organisation's standard supplements, such as IA-9 for service-to-service traffic.
- Assessment findings on tailored controls: findings raised against compensating controls, trended over time.
Continuous monitoring (CA-7 and CA-7(4) Risk Monitoring, both in all three baselines) is where these measures belong. See the NIST SP 800-137 Information Security Continuous Monitoring Guide and the NIST SP 800-53A Guide.
Framework mapping
| Framework | Relationship to SP 800-53B |
|---|---|
| NIST SP 800-37 Rev 2 (RMF) | Baselines and tailoring are the Select step; PL-2, PL-11 and CA-6 Authorization record the outcome |
| FIPS 199 and FIPS 200 | Categorisation and minimum security requirements that choose the baseline |
| FedRAMP | Low, Moderate, High and LI-SaaS baselines for cloud services, built on the Rev 5 baselines |
| NIST SP 800-171 Rev 3 | CUI requirements derived largely from the Moderate baseline |
| NIST CSF 2.0 | CSF outcomes map to 800-53 controls through NIST's informative references; a baseline is one way to reach a Target Profile |
| CNSSI 1253 | Equivalent baselines and overlays for national security systems |
Common failure patterns
- Tailoring by deletion: controls removed because they are hard, with no scoping consideration or compensating control. NIST is explicit that organisations do not arbitrarily remove controls.
- Treating the baseline as complete: the system faces insider or advanced threats that the baseline assumptions exclude, and nothing is supplemented.
- Stale inheritance: common controls inherited from a provider whose authorisation or scope changed.
- Blank parameters: organisation-defined values left unset, so assessors cannot test the control.
- Human-only identity scope: AC-2 and IA controls evidenced for users only, with service accounts and keys outside scope and IA-9 never added.
- Spreadsheet baselines: controls copied once and never compared with later NIST releases or overlays.
Practitioner checklist
- Record the FIPS 199 categorisation and the high-water mark decision with approval.
- Import the correct OSCAL baseline profile rather than retyping controls.
- List common controls and verify each provider's current authorisation.
- Set every organisation-defined parameter, starting from any overlay or FedRAMP values that apply.
- Record a scoping consideration type for each scoped-out control.
- Get compensating controls and residual risk agreed with the authorising official before build.
- Supplement for threats outside the baseline assumptions.
- Add IA-9 and NHI-specific specification text for any system with service-to-service traffic.
- Keep tailoring decisions under version control and re-review them on major change.
Standards and references
- NIST: SP 800-53B Control Baselines for Information Systems and Organizations
- NIST: SP 800-53B full text
- NIST: OSCAL content repository (SP 800-53 Rev 5 baseline profiles)
- NIST: SP 800-53 Rev. 5
- NIST: Security and Privacy Control Overlay Repository
- FedRAMP: Baselines
Related NHI Mgmt Group resources: NIST SP 800-53 Guide · NIST Control Overlays Guide · FedRAMP Guide · NIST Risk Management Framework (SP 800-37) Guide · NIST CSF 2.0 Govern Function Guide