Data Anonymisation Services That Hold Up Under UK GDPR
Anonymised data is information that can no longer identify anyone, by any means reasonably likely to be used. Get it right, with data anonymisation services that test the result, and UK GDPR stops applying to that dataset. Get it wrong and you’re still holding personal data, without the safeguards it needs.
Our data anonymisation services test each dataset, apply the right techniques and document the result. You can then share, analyse and keep data on firm legal ground. We work with privacy leads, data teams and engineers in finance, healthcare, telecom and SaaS.
What Is Anonymised Data and Why Do Businesses Need It?
Anonymised data means information that no one, including you, can use to identify a person by any means reasonably likely to be used. That test covers the data itself, any other information available, and the time, cost and technology someone would need.
Personal data comes with limits. You can’t keep it forever, share it freely or reuse it for any purpose you like. Properly anonymised information falls outside those rules.
That makes the anonymisation of data the safest way to analyse trends, share with partners, test systems and keep records beyond their retention period. The catch is that plenty of datasets labelled “anonymous” don’t meet the test.
Why Businesses Invest in Data Anonymisation Services
Each benefit of data anonymisation services comes from the same source: data that can no longer harm the people it came from.
Share Data Without Exposure
Partners and researchers get record-level insight. The shared set no longer identifies anyone, so your customers stay protected.
Keep Data Past Its Retention Date
Anonymised records can stay for trend analysis after personal data has to go. You keep the history without breaking the storage limitation principle.
Safer Test Environments
Engineers work with realistic data instead of production copies. Fewer people can see real customer records.
Answers for Due Diligence
Documented assessments show clients and auditors why a dataset counts as anonymous. You no longer rely on “we removed the names”.
Who Our Data Anonymisation Services Are For
Our data anonymisation services suit organisations that hold sensitive data and need to use it beyond its original purpose:
- DPOs and privacy leads who need a defensible position on what counts as anonymous
- Data and analytics managers sharing datasets with partners or research teams
- Engineering leads replacing production copies in test and staging
- Clinical and academic researchers preparing datasets for ethics review
- Founders of SaaS, fintech and healthtech firms facing vendor due diligence
Core Anonymisation Concepts
Anonymising personal data means removing these risks to a reasonable standard, not just deleting names.
Singling Out
Isolating one person’s record in a dataset, even when no name appears.
Linkability
Connecting records about the same person across two or more datasets.
Inference
Deducing something about a person with high probability from the data that remains.
Motivated Intruder
A capable person with no special access who sets out to re-identify someone. The ICO uses this as its starting test.
Is Anonymised Data Personal Data Under UK GDPR?
No, as long as the anonymisation holds. Recital 26 of UK GDPR says data protection law doesn’t apply to information that no longer relates to an identifiable person. The question is whether identification is reasonably likely, given the cost, time and technology available.
The ICO’s March 2025 anonymisation guidance treats identifiability as a spectrum. It doesn’t demand zero risk. It asks whether re-identification is reasonably likely for anyone who might hold the data.
Two points catch businesses out. The ICO treats anonymising personal data as processing in its own right, so you need a lawful basis and a clear purpose. The assessment also isn’t permanent, because new datasets and tools raise the risk over time. GDPR anonymised data needs a review date.
The Difference Between Anonymised and Pseudonymised Data
Anonymised data can’t be linked back to a person by any reasonably likely means. Pseudonymised data can, because someone holds the key or lookup table that reverses it. UK GDPR treats pseudonymised data as personal data, and anonymised information as outside its scope.
| Anonymised data | Pseudonymised data | |
|---|---|---|
| Can it be reversed? | No, by any reasonably likely means | Yes, with additional information kept separately |
| Personal data under UK GDPR? | No | Yes, for anyone who can re-identify it |
| Examples | Aggregated statistics, generalised or randomised datasets | Hashed emails, tokenised IDs, encrypted identifiers |
| Use it when | You never need to return to the individual | You need to re-link records later |
EU law added a nuance in 2025. In EDPS v SRB (September 2025), the Court of Justice held that pseudonymised data may not be personal data for a recipient with no realistic means of re-identification. For the organisation holding the key, it stays personal. The ruling isn’t binding in the UK, but it matters if you hold EU residents’ data.
When re-linking matters more than anonymity, our data pseudonymisation services are the better fit.
Data Anonymisation Techniques Explained
Real anonymisation combines techniques. No single method suits every dataset, and the ICO groups the main approaches into generalisation and randomisation.
| Technique | What it does | Best for |
|---|---|---|
| Generalisation | Replaces exact values with ranges or categories, such as age bands or partial postcodes | Reports and research datasets |
| Suppression | Removes rare values or records that would single someone out | Small groups and outliers |
| Aggregation | Publishes totals and averages instead of individual records | Dashboards and public statistics |
| K-anonymity | Makes every record match at least k−1 others on its identifying fields | Record-level data sharing |
| Noise and differential privacy | Adds calibrated randomness so no single record changes the result much | Published statistics and analytics |
| Synthetic data | Generates artificial records that follow the patterns of the original | Testing and model development |
Three methods get mistaken for anonymisation all the time. Hashing, encryption and tokenisation all leave a route back to the original, so they count as pseudonymisation. They still have their place. Data encryption protects data in storage and transit, data tokenisation protects payment and identity values, and data masking hides values in test and support systems.
Levels of Anonymisation: Where to Apply It
You can anonymise data at different points in its life, and our data anonymisation services cover each one. The right level depends on who needs the data and in what form.
| Level | Where it happens | Example |
|---|---|---|
| Field level | Individual columns at the point of export | Generalising dates of birth before a data share |
| Dataset level | A whole extract before it leaves the source system | A monthly research file for a university partner |
| Environment level | Every copy outside production | Staging and test databases |
| Output level | Results rather than records | Reports, dashboards and published statistics |
Output-level anonymisation carries the lowest risk. Record-level sharing carries the highest, so it needs the most testing.
Anonymising Data: Our Six-Step Process
Here’s how to anonymise data in a way you can defend. Each step produces a record you keep.
Discover and classify
We map where personal data lives and label every field as a direct identifier, indirect identifier or sensitive attribute. Our data discovery work feeds this step.
Assess re-identification risk
We test the dataset for singling out, linkability and inference, starting with the motivated intruder test.
Choose techniques
We match generalisation, suppression, noise or synthetic data to how you’ll use the data, so it stays useful.
Apply and test
We anonymise the agreed datasets, then try to re-identify them ourselves, including by linking them with public data.
Document the decision
You receive an assessment report covering the data, the techniques, the remaining risk and our reasoning.
Review on a schedule
We set a review date, because new data and tools change the risk. Where the processing is high risk, we connect the work to your DPIA through our DPIA services.
Your first call is free. We quote the assessment and anonymisation work once we’ve agreed which datasets are involved.
Regulatory Compliance and Frameworks
For anonymised data, GDPR in both the UK and EU sets the strictest test. The other frameworks below use different standards, and meeting one doesn’t mean you meet the others.
| Framework | How it treats anonymised information |
|---|---|
| UK GDPR and Data Protection Act 2018 | Outside scope if re-identification isn’t reasonably likely. The ICO’s guidance is under review following the Data (Use and Access) Act 2025. |
| EU GDPR | Same core test. The EDPB’s draft anonymisation guidelines (July 2026) add three criteria: no record isolation, no linkage and no inference. |
| HIPAA | Health data is de-identified under the Safe Harbor or Expert Determination method. HIPAA de-identified data can still be personal data under GDPR. |
| CCPA | Deidentified information is excluded only if you take reasonable measures, publicly commit not to re-identify it, and bind recipients by contract. |
Data protection anonymisation has to meet the strictest standard that applies to your data, not the most convenient one.
What Anonymisation Protects Against (and What It Doesn’t)
Anonymisation is strong protection, but it isn’t a cure-all.
Protects against
- Exposure of individuals if a shared or stolen dataset leaks
- Use of records beyond their original purpose
Doesn’t protect against
- Re-identification when weak techniques meet rich outside data
- Unfair conclusions about groups, even when no individual is named
AI, Synthetic Data and the Future of Re-identification
Re-identification keeps getting easier. A 2019 Nature Communications study estimated that 99.98% of Americans could be correctly re-identified in any dataset using 15 demographic attributes.
AI raises the stakes again. The EDPB’s draft guidelines say synthetic datasets and trained models aren’t automatically anonymous, because they can leak details through inference. Before you train on anything labelled as GDPR anonymised personal data, we test for memorisation and inference. Our AI and LLM data protection work covers the wider model risks.
Where Anonymisation of Data Creates Value
Anonymisation pays off wherever information needs to travel further, or last longer, than its original purpose allows.
Clinical Data Anonymisation Services
Research teams need patient-level data without patient identities. Our clinical data anonymisation services prepare datasets for ethics review and data sharing agreements, alongside our wider healthcare data protection work.
Learn moreFinance, Telecom and SaaS
Fraud models, network analytics and product usage data all gain from anonymised extracts. We support financial services firms, providers that need telecom data protection, and SaaS teams replacing production data in staging.
Why KEWDATA for Data Anonymisation Services
Legal advice alone won’t make a dataset anonymous. We work in the data as well as the law.
Legal and Technical in One Team
Our work is led by a former CISO and data protection adviser from UBS and Credit Suisse, so our assessments stand up with regulators and engineers alike.
Platform Experience
We work with privacy platforms including BigID, SecuPi, Securiti and PKWARE.
Multi-Jurisdiction Coverage
We cover UK GDPR, EU GDPR, HIPAA and CCPA from one provider.
Rated 5.0 on G2
Our rating is based on 8 verified reviews, several of which mention our anonymisation and tokenisation work.
If you need a wider check of your data protection position first, start with a GDPR audit.
What to Expect When You Work With Us
These are the questions buyers raise most, answered up front.
Cost
The first call is free. We quote the assessment and anonymisation work once we’ve agreed the datasets, with VAT shown.
Data That Stays Useful
We agree what analysis the data must support before we choose techniques, then test that it still does.
Your Existing Tools
We work with your databases, warehouses and pipelines, and with the privacy platforms you already run.
Ongoing Support
Re-identification risk changes, so we offer scheduled reviews. Hands-on changes inside your systems are scoped and quoted separately.
Data Anonymisation Checklist
Use these seven checks to see whether a dataset is ready to call anonymous.
- Every field classified as a direct identifier, indirect identifier or sensitive attribute
- Re-identification tested for singling out, linkability and inference
- Hashed, encrypted and tokenised fields treated as pseudonymised, not anonymised
- Small groups suppressed or merged before release
- Free-text fields reviewed for names and other details
- Lawful basis recorded for the anonymisation itself
- Assessment documented, with a review date
Frequently Asked Questions
Glossary of Key Terms
Anonymised Data
Information that can’t identify a person by any means reasonably likely to be used.
Pseudonymised Data
Personal data with identifiers replaced, which can be re-linked using information kept separately.
Motivated Intruder
A capable person with no special access who tries to re-identify someone.
Singling Out
Isolating one person’s record within a dataset.
Linkability
Connecting records about one person across datasets.
K-Anonymity
A standard where each record matches at least k−1 others on its identifying fields.
Differential Privacy
A method that adds calibrated noise so individual records barely affect the results.
Tom S.
Medical Practice
"Improved control over sensitive medical data"
"Practical guidance on access controls and data protection. Helped define structured access policies and introduced tokenisation for patient identifiers used in analytics."
Sarah S.
Retail, Enterprise
"Reliable partner for telecom data protection"
"Quickly understood telecom data complexity; helped classify sensitive datasets and apply protection measures for subscriber, usage, and billing data."
Pauliina H.
Entertainment, Enterprise
"Effective support for securing customer data"
"Identified where sensitive data lived and applied tokenization and anonymization strategies in a complex telecom environment."
Jordan R.
Computer Networking
"Practical approach to safeguarding e commerce"
"Helped protect PII, payment data, and order histories with tokenization and anonymization; advised on GDPR compliance for international operations."
Charmaine S.
Int. Trade & Development
"Systematic Approach Enhances Data Security"
"Methodical approach through discovery, planning, and implementation; implemented Microsoft Purview for a scalable compliance framework."
Kateryna H.
IT and Services
"Practical approach to protecting sensitive Data"
"Structured data discovery and classification combining technical analysis with finance data privacy expertise; strong masking and tokenization policies."
Snow D.
Market Research
"Practical expertise for telecom information"
"Introduced tokenization and controlled access strategies for subscriber data while still supporting analytics and reporting."
Laura H.
Hospital & Health Care
"Valuable expertise in healthcare protection"
"Strong expertise in healthcare data security; introduced anonymization techniques for safely using patient data in research and reporting."
Make Your Anonymised Data Defensible
Find out whether the data you call anonymous would pass the ICO’s test. Book a free call, and we’ll review your datasets, show you where the risks are, and quote the work to fix them.
Book a free anonymisation risk reviewYour first step is a free call about the data you hold.
