Privacy by Design - Building Privacy In from the Start
About 2 min read
Privacy by Design is a design principle that builds privacy protection into systems and services from the planning and design stages. Rather than "adding privacy measures afterward," it places privacy at the very core of the architecture, structurally reducing the risk of leaking personal information. It was advocated in the 1990s by Dr. Ann Cavoukian, then the Information and Privacy Commissioner of Ontario, Canada, and was codified as a legal obligation in Article 25 of the GDPR, which took effect in 2018.
Dr. Ann Cavoukian's 7 Principles
Prevent privacy violations from occurring rather than responding after they happen.
Ensure that privacy is protected even when the user configures nothing.
Privacy is not an add-on feature but is integrated into the core design of the system.
Privacy and functionality are not a trade-off; both can be achieved together.
Ensure privacy at every stage, from data collection to disposal.
Disclose how processing works and enable independent verification.
Prioritize the interests of individuals and let users control their own data.
Legal Obligation Under GDPR Article 25
Article 25 of the GDPR, titled "Data protection by design and by default," elevated Privacy by Design to a legal obligation. Data controllers must implement appropriate technical and organizational measures to apply principles such as data minimization, both at the time of determining the means of processing and at the time of processing itself. Violations may incur fines of up to 10 million euros or, for an undertaking, up to 2% of total worldwide annual turnover of the preceding financial year, whichever is higher.
The Principle of Data Minimization
Article 25 of the GDPR names data minimization as an example of the data-protection principles that the measures are to implement. Data minimization is the principle of "collecting and retaining only the minimum data necessary for the purpose," and concretely it means design decisions such as the following. Do not collect a real name if it is unnecessary for user registration. Anonymize data if there is no need to identify individuals for analysis. Set retention periods and automatically delete data that is no longer needed. Use data masking and pseudonymization to keep access to raw data to a minimum.
Privacy by Default
- All data public by default
- Opt-out model (data is shared unless you change the setting yourself)
- Continuously collect location data
- Retain data indefinitely even after account deletion
- Private by default
- Opt-in model (shared only when explicitly consented)
- Collect location data only when needed
- Set retention periods and delete automatically
Contrast with Dark Patterns
Dark patterns are a design technique at the opposite extreme of Privacy by Design. Techniques such as making the consent button prominent while shrinking the reject button, complicating the settings screen to make changing privacy settings difficult, and deliberately making the account-deletion process cumbersome all impede users' autonomous decision-making. The EU Digital Services Act (DSA) has prohibited providers of online platforms from interface designs that deceive or manipulate users (dark patterns) since 2024.
Why the Design Stage Costs Less Than Retrofitting
The from-the-design-stage part of Privacy by Design is not mere idealism; it describes a cost structure. If the question of whether a piece of data is truly necessary for the purpose is raised at design review, changes such as anonymizing identifiers or trimming log fields amount to no more than revising the blueprint. Performing the same review after a service has launched requires migrating or reworking accumulated data and checking the impact on existing features, inflating the scale of the work. Moreover, data that was never collected generates no protection, auditing, or deletion work. Because the burden of regulatory compliance grows with the volume and sensitivity of the data held, minimization at the design stage keeps paying off as a lasting reduction in compliance workload.
For balancing privacy and convenience, see the article on privacy versus convenience; for specific configuration methods, see the privacy settings guide; and for protecting your digital identity, see the article on digital identity protection, which explain these topics in detail.
Was this article helpful?
Related Terms
- Pretexting - Fabricated Scenarios for Data TheftPretexting is a social-engineering technique that fabricates a scenario to extract information. Impersonating IT support or HR is typical, and it often serves as the opening move in a BEC attack.
- Cyber Kill Chain - Seven Phases of an AttackThe Cyber Kill Chain is a framework that breaks cyberattacks into seven phases. Proposed by Lockheed Martin in 2011, it helps defenders systematically plan countermeasures at each stage.
- OSINT - Open Source IntelligenceOSINT gathers intelligence from publicly available sources such as social media, WHOIS, and DNS records. Both attackers in reconnaissance and defenders in threat analysis use it.
- BEC - Business Email CompromiseBusiness email compromise impersonates executives or partners via email to trick targets into wire transfers or sharing confidential data. FBI reports put cumulative losses above $50 billion, with AI deepfakes raising the threat.