A secure configuration baseline is a written, agreed set of settings that every device, server, cloud service or network appliance of a given type must meet before it goes into use. You build it by choosing a trusted starting point, removing what you do not need, locking down accounts and access, recording the result, and then checking regularly that nothing has quietly moved away from it. This guide shows how to do that in a small or mid-sized UK organisation, with a practical hardening checklist and a note on where it meets Cyber Essentials.
Key takeaways
- A baseline is a documented, approved configuration that changes only through change control. A secure baseline is the version that represents the most secure state your operations can live with.
- Start from vendor and NCSC guidance rather than a blank page, then record the exceptions you genuinely need.
- The core hardening moves are the same everywhere: remove unused software and accounts, change default passwords, limit admin rights, switch off unused ports, and keep everything patched.
- A baseline only helps if you check against it. Drift is the normal state of any system nobody is watching.
- Secure configuration is one of the five Cyber Essentials technical controls, so a good baseline is also most of your evidence for certification.
Table of Contents
- What a secure configuration baseline actually is
- Why bother: what a baseline removes
- How to build a baseline in six steps
- System hardening checklist
- Keeping the baseline true: drift and review
- How a baseline supports Cyber Essentials
- Common mistakes
- Testing whether your baseline holds
- Getting help from Solusec
- FAQ
- Sources
What a secure configuration baseline actually is
The terms get mixed up, so it helps to separate them. NIST defines a baseline configuration as a formally reviewed and agreed set of specifications for a system, which can then be changed only through organisational change control procedures. The same NIST bulletin describes a secure baseline as one developed and approved to represent the most secure state consistent with operational requirements and constraints.
In plain terms:
- Configuration baseline: the agreed standard build, whether or not it is especially secure.
- Security baseline: a set of recommended security settings, often published by a vendor for its own products.
- Secure configuration baseline: your own approved standard, built from those recommendations and adjusted for how your business works.
People also call it a gold build, a standard build or a hardened image. The label matters less than the habits: it is written down, someone approved it, and deviations are noticed.
Why bother: what a baseline removes
Most breaches of small organisations do not need clever attacks. They use a default password, an old service nobody remembered to switch off, or an admin account used for email. A baseline takes those choices away from individual people.
UK government guidance has long framed it this way: organisations need secure baselines to remove or disable unnecessary functionality and keep systems patched, which reduces exposure to vulnerabilities. That page has since been withdrawn, so treat it as background, but the principle is carried into current requirements. The Cyber Essentials requirements say organisations need secure configuration to reduce vulnerabilities and make sure devices provide only the services needed for their role.
There are three practical benefits beyond security:
- Faster setup. A new laptop or server built from a baseline is ready in minutes and identical to the last one.
- Easier support. When every machine of a type is the same, faults are easier to reproduce.
- Cleaner evidence. When a customer, insurer or auditor asks how you configure devices, you hand over a document rather than an explanation.
How to build a baseline in six steps
1. Inventory what you have
You cannot baseline what you cannot list. Record each class of asset: Windows laptops, Macs, phones, servers, firewalls and routers, Wi-Fi access points, cloud tenants (Microsoft 365, Google Workspace, AWS, Azure) and any business applications. For a 30-person firm this is a one-page spreadsheet. Group by type, because you will write one baseline per type, not one per device.

2. Pick a trusted starting point
Do not invent settings. For user devices, the NCSC maintains guides on how to securely configure each of the platforms commonly used in the UK. Microsoft also publishes security baselines for Windows and its other products, and for servers and network kit you should use the vendor's own hardening guide. Where you want a more detailed benchmark, industry consensus benchmarks exist for most operating systems, though a vendor or NCSC guide is a perfectly good place for a small team to begin.
3. Decide what to change, and what to leave
Every hardened setting has a cost in convenience. Walk through the starting point and mark each setting as adopt, adopt with a change, or reject with a reason. The reasons are the valuable part: "Legacy accounting package needs this protocol, compensating control is network segmentation" is a perfectly good entry. A baseline with documented exceptions is stronger than a perfect-looking one nobody follows.
4. Test on a pilot group
Apply the baseline to two or three real devices used by people who will tell you when something breaks. Fix what needs fixing before you roll it to everyone. Settings that block a finance team's macro-enabled spreadsheet will be found within a day.
5. Write it down and get it approved
The document can be short: scope, the settings, exceptions, the owner, the date approved and the date of next review. Approval by a director or the IT manager is what makes it a baseline rather than a suggestion. Store the build itself (image, policy export or script) alongside the document.
6. Build new devices from it, and change it only deliberately
Any change to the baseline goes through the same approval, with a version number. This is what the NIST definition means by change control, and it is what stops the baseline becoming a record of what someone did last spring.
System hardening checklist
The withdrawn 10 Steps guidance gave a short list that is still sound: remove unnecessary services, disable unused ports including USB where appropriate, limit user rights and keep systems patched. Here it is expanded into a working checklist by asset type.
| Area | Baseline settings to check |
|---|---|
| Operating system (laptops, desktops, servers) | Supported version only; automatic updates on; built-in firewall on; disk encryption on; screen lock after a short idle period; no local admin rights for everyday users |
| Accounts | Unused and leaver accounts removed or disabled; default passwords changed; separate admin accounts used only for admin tasks; multi-factor authentication on cloud services |
| Software | Unneeded applications and browser extensions removed; only approved software installed; licences and versions recorded |
| Ports and peripherals | Unused ports and services disabled; USB storage restricted where there is no business need; remote desktop not exposed to the internet |
| Network devices | Default credentials changed; admin interfaces not reachable from the internet; firmware current; unused features switched off; logging enabled |
| Mobile devices | PIN or biometric lock; encryption; ability to wipe remotely; OS updates applied |
| Cloud services | Tenant-wide MFA; admin roles minimal; public sharing and storage access reviewed; logging turned on; settings reviewed against the provider's good-practice guide |
A few of these deserve a note.
Network equipment. The NCSC expects vendors to make network products secure by default and to document how to configure, manage and update them securely. That helps you, but you still have to apply it: change the default admin credentials and check that the management page is not open to the internet. The NCSC also encourages manufacturers to enable robust logging by default, so confirm logging is actually on and kept somewhere useful.
Cloud. Secure-by-default options do not remove your responsibility. NCSC cloud guidance says the customer remains responsible for configuring the service, and that organisations should periodically review their configuration against the provider's good practice guides. Cloud misconfiguration is one of the most common findings we see in assessments, which is why we offer a dedicated cloud security assessment for AWS and Azure.
Keeping the baseline true: drift and review
A baseline is a snapshot. Within months, people install tools, grant temporary access that becomes permanent, and fix urgent problems with a one-off change. That gap between the approved state and the real state is configuration drift.
The NCSC Cyber Assessment Framework describes the expected habits: keep a record of the current known-good configuration, closely manage and document changes, and regularly review and validate systems against their expected secure settings. For a small organisation that translates to:
- Quarterly spot checks. Pick five devices at random and compare them to the baseline document. Record what you find.
- Policy-based management. If you use Microsoft 365 or an MDM tool, enforce settings centrally so a device that drifts is corrected or flagged automatically.
- Change log. Every change to a server, firewall or cloud tenant is noted with who, what and why. A shared document is enough to begin with.
- Review on trigger. Revisit the baseline when a vendor releases new guidance, when you adopt a new platform and after any incident.
Cloud and infrastructure as code
If you run cloud infrastructure, define it in code. The NCSC notes that using infrastructure as code and deploying through pipelines means manual changes are avoided, which helps prevent drift from template configurations. Cloud security posture management tools can also monitor environments for misconfigurations and drift from established baselines. The same NCSC piece is candid that they are not a cure-all: behaviour that looks wrong in one workspace may be needed in another, so a person still has to judge the findings.
Watch for the unusual
Configuration checks tell you what is set. Monitoring tells you what is happening. The ICO advises organisations to establish a baseline for normal system activity and monitor against it to spot unauthorised access or misuse. If you do not have the people to watch logs, a SOC consulting and support arrangement can cover that gap.
How a baseline supports Cyber Essentials
Secure configuration is one of the five technical controls in Cyber Essentials. The government describes it as one of five key protections businesses can use to show customers and suppliers they take cyber security seriously.
Under the April 2026 requirements (v3.3), you must meet every requirement that applies to your scope, including regularly removing or disabling unnecessary user accounts and software, changing default or guessable passwords and applying device-locking controls. Those items sit directly in the checklist above, so a baseline built this way gives you most of your evidence.
Two cautions. First, a baseline does not replace the requirements: Cyber Essentials also covers firewalls, user access control, malware protection and security update management, and you answer the questionnaire against the current version. Second, the scheme changed this year, so read our summary of the April 2026 Cyber Essentials changes and the note on why MFA is now a hard fail. Our Cyber Essentials checklist turns the requirements into a pre-submission review, and how to get Cyber Essentials walks through the process. Schools and colleges have their own pressures, covered in our guide to Cyber Essentials for schools.
Common mistakes
- Copying a benchmark wholesale. Applying every setting without testing breaks business software, and people then bypass the controls. Adopt, adjust, document.
- One baseline for everything. A finance laptop, a kiosk PC and a build server need different settings.
- No owner. An unowned baseline is out of date within a year.
- Forgetting the edge. Routers, printers, cameras and access points are often left on factory settings.
- Treating the document as the control. The control is the enforced setting and the check. The document is the record.
- Assuming the cloud is configured for you. It is not. See the NCSC responsibility point above.
Testing whether your baseline holds
Checking against your own document proves consistency, not strength. A baseline can be followed perfectly and still leave an exposed service or a weak web configuration. An outside test tells you what an attacker would find. Our published prices are £3,000 for an external IP test and £2,500 for a CMS website test, both delivered by CREST-accredited testers. If you are unsure whether you need a scan or a test, read penetration test vs vulnerability scan.
Getting help from Solusec
If you want the baseline built and the evidence checked by people who do this every week, we can help. For certification, our Cyber Essentials service is priced at £320 to £600 depending on organisation size, a gap analysis before you submit is £300, and urgent turnaround is an extra £100. You deal directly with the assessor, not a sales team, and the findings come in plain English with proportionate recommendations. If you need stronger assurance, we also assess Cyber Essentials Plus and IASME Cyber Assurance.

For a first conversation about your configuration, certification or testing, start at solusec.co.uk.
FAQ
What is a secure baseline configuration?
It is a documented, approved set of security settings that a type of system must meet before it is used. It represents the most secure state your operations can support, and it changes only through a controlled, recorded process.
What is the difference between a security baseline and a configuration baseline?
A configuration baseline is any agreed standard build. A security baseline is a set of recommended security settings, usually from a vendor. A secure configuration baseline combines the two: your approved build, hardened using those recommendations.
How often should I review a baseline?
Review it at least yearly, and whenever you adopt a new platform, a vendor publishes new guidance, the Cyber Essentials requirements change or you have an incident. Check real devices against it more often, quarterly for small organisations is a sensible rhythm.
Do I need special tools to keep one?
Not to begin with. A written document, a standard build and a regular spot check are enough for a small team. Central management through your device or cloud platform becomes worthwhile as you grow, and posture management tools help in larger cloud estates.
Sources
- NIST ITL Bulletin: Managing the Configuration of Information Systems
- 10 Steps: Secure Configuration, GOV.UK (withdrawn)
- Cyber Essentials Requirements for IT Infrastructure v3.3, NCSC
- NCSC device security guidance: platform guides
- NCSC Vendor security assessment
- NCSC guidance on digital forensics and protective monitoring specifications
- NCSC Cloud security principle 14: secure use of the service
- NCSC Cyber Assessment Framework, principle B4 system security
- NCSC: Security benefits of a good cloud service
- NCSC: Cloud Security Posture Management blog
- ICO: System and network security
- GOV.UK: Businesses urged to lock the door on cyber criminals
Recommended
- How to Get Cyber Essentials | Step by Step
- Cyber Essentials Checklist: What You Need to Pass (2026)
- Cyber Essentials v3.3: MFA is now a hard fail
- Cyber Essentials April 2026 Changes (v3.3 / Danzell)
This article was created with assistance of AI. We never use AI with your data unless given explicit permission.
