ISO 27001 Control A.5.13: Labelling of Information
Classifying information is a good first step, but a classification that lives only in a policy document does not protect anything. Someone has to look at a spreadsheet, an email or a USB drive and know, within a second or two, how careful to be with it. A.5.13 is the control that makes classification visible. It is the 13th of 37 organizational controls in our walk through Annex A.
Labelling sounds trivial until you try it across a real organization. Documents are shared as PDFs, copied into chat messages, exported from databases and printed for meetings. A label that works on a Word file may vanish the moment the content is pasted elsewhere. The goal of this control is not a perfect stamp on every byte. It is a practical, repeatable way of signalling handling requirements to the people and systems that touch the information.
What the control requires
Annex A control 5.13 of ISO 27001:2022 states that an appropriate set of procedures for information labelling shall be developed and implemented in accordance with the information classification scheme adopted by the organization.
The key elements:
- Appropriate set of procedures - written, proportionate steps for how labels are applied, not a vague instruction to “mark things confidential”
- Information labelling - covering information in all its forms, including documents, records, emails, files, storage media and, where it makes sense, physical items
- In accordance with the classification scheme - the labels must mirror the levels you defined, so the two controls work as a pair
- Developed and implemented - the procedure has to exist on paper and be followed in practice
A.5.13 is a preventive control in ISO/IEC 27002:2022. It stops mishandling before it happens by telling people what they are holding. Annex A states the control; ISO 27002 gives guidance, including the point that labelling needs to be easily recognizable, and that information in different formats may need different labelling methods. It also notes that labelling can have drawbacks: labelled information is easy to identify, which is useful to an attacker looking for the valuable material. That is a reason to think about metadata and access control, not a reason to skip labels.
How to implement labelling of information
Start from the classification scheme you already have
Labelling follows classification of information. If your scheme has four levels, such as public, internal, confidential and restricted, you need a label for each, and everyone must understand what each one means. Resist the urge to invent extra labels at this stage. Every new label is another thing people will get wrong. Keep the wording identical to the scheme so that a person can move from the policy to a document without translating.
It helps to write a short table that maps each level to a label text, a color or marking if you use one, and the handling rules that follow from it. One page is enough. Put it where people actually look, such as the intranet page for the information security policy and the template library.
Decide how each format is labelled
Different formats need different methods. A practical approach is to list the formats your organization really uses and decide a method for each:
- Documents and presentations - a header or footer with the label, applied from the template so authors do not have to remember
- Email - a prefix in the subject line or a banner added by the mail system, plus a default for messages that contain nothing sensitive
- Files and folders in shared storage - sensitivity labels or tags in your document platform, or folder naming conventions where tooling is limited
- Databases and applications - labels as metadata on records or tables, or an entry in your inventory of information and other associated assets that names the classification of the whole system
- Physical media and printouts - a printed or handwritten marker on the item or its cover, as well as on storage media such as encrypted drives and backup tapes
Where you can automate, do. Templates with built-in labels and tools that apply labels by default are far more reliable than a rule asking people to type the right word. Where you cannot automate, make the manual step small.
Handle exceptions and unlabelled information
Some information cannot reasonably carry a label: a log stream, a machine-to-machine feed, a conversation. The procedure should say what happens in those cases. Typically the system as a whole is classified and handled accordingly, and the label lives in the inventory or the system documentation rather than on each item.
Decide what unlabelled means. A sensible default is that unlabelled information is treated as internal until the owner says otherwise, and that anything received from outside keeps the label of the sender where one exists, or is classified on arrival. Whatever you choose, state it plainly. Auditors will ask what a person should do when they find a file with no label.
Also define what happens when information is combined. A report that pulls together internal figures and one confidential table should usually take the higher label. Say so in the procedure, because it is the situation people argue about most.
Make handling rules follow the label
A label with no consequence is just decoration. Link each label to concrete behavior: who may receive the information, how it may be sent, whether it can be printed, how it must be stored and how it is destroyed. The acceptable use of information and other associated assets rules are the natural home for these. Rules for sending information outside the organization belong in your procedures for information transfer, which we cover next in the series.
Finally, train people. Short, concrete examples work better than definitions: “this contract draft is confidential, so it goes in the restricted folder and gets the footer from the template”. Cover labelling in information security awareness, education and training and in onboarding, and revisit it when you change tooling.
Related controls and clauses
A.5.12 connection. Classification of information defines the levels that labels must reflect. Without it there is nothing to label.
A.5.9 connection. The inventory of information and other associated assets is a convenient place to record the classification of systems that cannot carry visible labels.
A.5.10 connection. Acceptable use of information and other associated assets turns labels into handling rules people can follow.
A.7.7 connection. Clear desk and clear screen rules are easier to apply when printed material carries a visible label.
A.7.10 connection. Storage media needs labelling so that drives and tapes are handled and disposed of according to what they hold.
A.6.3 connection. Information security awareness, education and training is how people learn what each label means.
Clause 7.5.3 connection. Clause 7.5.3, Control of documented information, already expects you to identify and control your ISMS documents, so labelling those is a good place to begin.
Forward reference. Information transfer, the next control in the series, uses labels to decide how information may leave the organization. We cover it as A.5.14.
What auditors check
A written procedure. Auditors look for a documented labelling procedure that matches the classification scheme level for level. A procedure that uses different words from the scheme is a quick finding.
Labels in real documents. They sample documents, emails, shared folders and media and compare what they see with the procedure. Expect them to ask to see a few classified items and to check the label is present and correct.
Coverage of formats. They ask how labelling works for systems, databases, physical media and verbal information, and whether the answers are consistent.
Handling that follows the label. They check that a confidential item is actually stored, sent and destroyed the way the rules say, not only that it carries a label.
Staff understanding. Interviews show whether people know the levels, can explain what a label means and know what to do with unlabelled items.
Treatment of exceptions. Auditors want to see that the default for unlabelled information and the rule for combined information are defined and known.
Common mistakes to avoid
Labelling everything the same. If every document is stamped confidential, the label carries no information and people stop noticing it. Use the lowest level that genuinely applies.
Too many labels. Seven levels with subtle differences produce inconsistent use. Three or four is usually plenty for organizations of our clients’ size.
Relying on memory. A policy telling people to add a footer will be followed for a week. Put the label in the template or the tool.
Labels without handling rules. If nobody can say what to do differently with a restricted item, the exercise is a checkbox. Tie each label to specific behavior.
Forgetting physical and removable items. The unlabelled USB drive in a desk drawer is a classic audit finding. Include media and printouts in the procedure.
Never reviewing. Information changes over time, and a label that was right at creation may be wrong a year later. Give owners a reason and a way to relabel.
How 27kay can help
We help organizations build labelling that people actually use, starting from a classification scheme that fits how you work rather than copying a template with seven levels. As part of ISO 27001 implementation, we review your documents, tooling and media, design simple labelling rules for each format and prepare the evidence an auditor will ask for. For the full picture, see our ISO 27001 knowledge hub.
If someone picked up your most sensitive file today, would they know from the file itself how careful to be? Get in touch - we will help you make sure the answer is yes.