What is Data Anonymization?

Data anonymization techniques are the modification of data in systems in such a way as to prevent the data from pointing to a specific individual while maintaining the format and consistency of the data. It is one of the approaches organizations can use to comply with strict data privacy laws that require the protection of personally identifiable information (PII), such as contact information, health records, or financial information.

Why Data Anonymization is Important

Even if they gain business value from it for customer support, analytical insights, test data, outsourcing purposes of supplier service, and more, data anonymization helps companies keep PIIs confidential by hiding sensitive information.

What Are the Key Benefits of Data Anonymization?

Data anonymization is a way to demonstrate that your company accepts and implements its responsibility to protect sensitive, personal and confidential data in an environment of increasingly complex data privacy directives that may vary depending on where you and your global customers are located.

Customers who entrust their sensitive information to companies will also consider a breach of this data as a breach of their trust in the company and as a result decide to seek services from a different company. An industry survey showed that 85% of consumers would not do business with that company if they were concerned about a company's privacy practices, and only 25% of respondents believed that many companies handle their PII correctly.

In addition to protecting companies against possible loss of trust and loss of market share, data anonymization is also a defense against the risks of data breaches and the risks of data misuse by company officials. The fine imposed in case of violation of the General Data Protection Regulation (GDPR) can be, for example, 10 million euros - 20 million euros, or 2-4% of the global annual turnover, whichever is more. Even a single complaint can lead to a costly and time-consuming audit.

But data anonymization is not just about avoiding risk — it also improves data governance and data quality. With clean, reliable data, you can optimize applications and resources, maintain big data privacy and analytics, and accelerate cloud workloads, all of which drive digital transformation by unlocking secure data to use in creating new business value.

What Data Should Be Anonymized?

Regardless of whether a company stores or processes PIIs about EU citizens, strict GDPR requirements provide a useful benchmark for the types of data to be protected. The GDPR defines personal information as “any information relating to an identified or identifiable data subject” and this information includes:

· Basic credentials such as name, address and identification numbers

· Web data such as location, IP address, cookie data and RFID tags

· Health and genetic information

· Biometric data

· Racial or ethnic data

· Political views

· Sexual orientation

Many companies must also comply with industry-specific regulations. Independence Health Group, a US health insurance company, is an example of how data anonymization should be successfully implemented for healthcare regulations. Independence Health Group is governed by HIPAA, which strictly regulates the processing of Americans' protected health care information (PHI). The company must protect the PHI of 8.3 million insured persons, both to avoid the high fines and corrections that would arise from the breach of health care data, and to protect the well-being and trust of its customers. But at the same time, the insurance company is required to cooperate with external data processing partners and allow developers, both internally and externally provided, to test applications on the relevant data.

To test high-quality applications and process data without the risk of unauthorized access, Independence Health Group uses Dynamic Data Masking to anonymize a lot of data, from names, birth dates, and Social Insurance Numbers to diagnostic and billing records.

Are There Alternatives to Data Anonymization?

Permanent data masking for anonymization

It can be used for data masking, anonymization or aliasing. It enables it to work with masked results by using characters that will typically maintain format requirements for an application, and replaces data elements with similar-looking representative elements. Permanent data masking is typically used for anonymization. Dynamic data masking is reversible and can transform data during operation, based on user role and context, to ensure real-time operational systems for more flexible data privacy and maintenance.

When data is masked, persistent data masking does not contain any reference to the original information and cannot be reversed, reducing the risk of potentially inappropriate data exposure. It is most commonly used for test data, research and development on highly sensitive data, or precision projects. Permanently masked data cannot be converted to unmasked.

Dynamic data masking for aliasing

Aliasing data is used to replace personal data fields in a record with alternative representative values. Giving a nickname does not remove all possible identity identifiers from the data and can be undone, so if you have additional information that can link or renew the alias given to the original data, there is the possibility of redefining it.

For example, if you have a dataset of employees' names, email addresses, phone numbers, and salaries, an attack to uncover patterns in these fields could discover the original values. Alternatively, simple access to encryption keys used to “unmask” pseudonymous data or similar data transformation controls can be used to return representative values to their completely unmasked original state.

Due to the possibility that the data can be reidentified directly or indirectly, aliasing should not be used in cases where you need a complete disconnect between the individual's identity and data—only data anonymization can completely hide data from potential determinants. Seen from the good side, aliasing presents a manageable risk when there are legitimate use cases for the data to be returned to the original values later.

Data Encryption

Data encryption is another form of data protection that uses algorithms to convert unencrypted texts into an illegible form, losing the original format of the data and making it unusable in the new state. Data encryption is useful for stationary and moving data, such as storage or network connections, where the use of data is not an urgent requirement. Unlike anonymization, data encryption is reversible; encrypted data can be retrieved by any person with the encryption key to the corresponding decryption algorithm. This necessitates the use of a complex encryption algorithm that cannot be easily broken, as well as protection against access to the keys to the data.

Encryption is commonly used to protect files in transit or static, but it offers flexibility when these files need to be used for later reidentification—for example, linking successful clinical trial results back to specific patients for future follow-up.