Data minimization is the practice of collecting, using and keeping only the personal data a task genuinely needs. If a field, a log or a sensor reading does not serve a stated purpose, the safest choice is not to gather it at all.
It is one of the most practical ideas in data privacy, because data that was never collected cannot be leaked, misused or subpoenaed later.
What data minimization means in practice
The idea is simple to state and harder to apply. Before an organisation, a developer or a hackathon team gathers personal data, it asks what the product has to do, which pieces of information that job requires, and how long each piece must stay around. Anything beyond those answers is surplus. Surplus data adds cost and risk without adding value.
Minimization has three working dimensions: how much data is collected, how widely it is used inside the organisation, and how long it is kept. A sign-up form that asks only for an email address is minimal on the first dimension. A database where only the support team can see that address is minimal on the second. A rule that deletes inactive accounts after a fixed period handles the third.
The principle also guards against a common habit: collecting everything because storage is cheap and the data might be useful someday. “Might be useful” is not a purpose. Regulators that enforce privacy laws generally expect a purpose to be defined before collection starts, and the data to be tied to it.
What the GDPR text says
The European Union’s General Data Protection Regulation puts minimization into law. Its principles article describes personal data that is “adequate, relevant and limited to what is necessary in relation to the purposes for which they are processed”, and labels this principle data minimisation. The same article adds a time limit, requiring that data is kept in a form which identifies people for no longer than is necessary for the purposes of the processing.
A later provision, on data protection by design and by default, turns the principle into a design duty. It says the controller must implement measures so that by default only personal data which are necessary for each specific purpose are processed. According to the regulation’s text, that obligation applies to the amount of personal data collected, the extent of their processing, the period of their storage and their accessibility.
| Dimension | What the text asks of a controller | Example of a minimal choice |
|---|---|---|
| Amount collected | Only data necessary for each specific purpose | A newsletter form that asks for an email address and nothing else |
| Extent of processing | Use data only as far as the stated purpose requires | Not feeding support tickets into unrelated marketing profiles |
| Period of storage | Keep identifiable data no longer than necessary | Deleting event sign-up data once the event is finished and any follow-up duty is met |
| Accessibility | Do not make data available to an indefinite number of people without the individual’s intervention | Role-based access so staff see only the records their job needs |
The first column follows the wording of the official GDPR text on EUR-Lex; the examples are illustrations, not legal requirements.
Design by default and the role of pseudonymisation
The design-and-default provision names pseudonymisation as an example of an appropriate measure. Pseudonymisation replaces direct identifiers with codes, so a record can still be used for analysis without showing whose it is. One of the regulation’s explanatory recitals says that applying pseudonymisation “can reduce the risks to the data subjects concerned”.
It is worth being precise about what this does and does not achieve. The regulation treats pseudonymised data that could be attributed to a person with additional information as still being information about an identifiable person. So pseudonymisation lowers risk but does not make data anonymous, and it is a complement to collecting less rather than a substitute.
The European Data Protection Board, which brings together the EU’s data protection authorities, has published guidelines on data protection by design and by default that explain how regulators read these duties. Organisations that need to apply them should read the current guidance directly, since regulators update their positions over time.

How the United States approaches minimization
Privacy rules in the United States come from sector laws and from state legislation, so the exact duties depend on where a business operates and whose data it handles. California’s law is a well-known example. The California Attorney General’s page says the CCPA “gives consumers more control over the personal information that businesses collect about them”, and that the CCPA regulations provide guidance on how to implement the law. That page was marked as updated on August 28, 2026.
Because state laws differ and change, anyone building a product for US users should check the current statute and regulations for each state, or ask a qualified lawyer. This guide describes the principle, not the obligations of any particular business.
Regardless of which law applies, the engineering habit is the same. A team that can say what each field is for, and that deletes what it no longer needs, will find compliance with any of these laws easier than a team that stores everything by default.
Practical steps for collecting less
Minimization works best as a routine rather than a one-off clean-up. The following steps suit small teams, civic projects and hackathon prototypes as well as larger organisations.
- Write the purpose first. For each data field, record in plain words why it exists. If nobody can answer, remove the field.
- Make optional fields truly optional. Asking for a phone number “just in case” turns a convenience into a liability.
- Prefer coarse data to precise data. A town or region is often enough where a street address is not needed, and an age band can replace a full birth date.
- Set retention rules. Decide at the outset when each category of data will be deleted, and automate the deletion.
- Limit access. Give people and services only the permissions their role needs, and log who looks at sensitive records.
- Aggregate early. If a dashboard needs counts, store counts, not the raw records that produced them.
- Review third-party tools. Analytics scripts, embedded widgets and plug-ins can collect data on your behalf; each one should justify itself.
The US National Institute of Standards and Technology offers a structured way to think about this. Its Privacy Framework is described as a voluntary tool intended to help organizations identify and manage privacy risk while building products and services. Teams that want a vocabulary for discussing data practices can use it as a checklist for conversations between engineers, lawyers and managers.
Minimization in civic and community projects
Community technology projects face a particular temptation. Residents are often willing to share information for a good cause, and the project team can end up holding records far beyond what the project needs. A neighbourhood survey that records names alongside opinions, or an app that logs precise locations by default, creates a store of sensitive data that the team may be poorly equipped to protect.
The same logic applies at city scale. Sensors and cameras can capture data about everyone who passes by, which is why the discussion of smart city privacy keeps returning to purpose limits, short retention and clear signage. A city that decides in advance what a sensor is for, and configures it to record only that, has applied minimization before a single reading is stored.
Volunteer-run projects have an additional reason to be strict. They often lack a security team, a legal department or a formal process for handling a breach. Holding less data is the cheapest security measure available to them.
What individuals can do
Minimization is usually presented as an organisational duty, but individuals can apply it from the other side. When a form asks for more than the service obviously needs, it is reasonable to leave optional fields blank, use a separate email address for sign-ups, or decline to create an account when guest access exists.
Readers who want a broader checklist can see how to protect your privacy online, which covers account settings, browsers and everyday habits. Both sides matter: organisations that collect less and individuals who share less together shrink the pool of data that can go wrong.
It also helps to remember that people have rights over the data an organisation does hold. The California Attorney General’s CCPA page lists a right to know about the personal information a business collects. How such rights work in a given place depends on the current law, so readers should check the regulator’s guidance for their jurisdiction.
Frequently asked questions
Is data minimization only a European idea?
No. The principle is written into the GDPR, but the underlying idea of collecting no more than needed appears in privacy frameworks and state laws elsewhere, including guidance from NIST in the United States. The legal wording and enforcement differ by place.
Does data minimization mean never collecting personal data?
No. It means collecting what a defined purpose requires and no more. A shop needs a delivery address to deliver goods; it does not need that customer’s date of birth.
Is pseudonymised data the same as anonymous data?
No. Under the GDPR, pseudonymised data that could be linked back to a person with additional information is still treated as information about an identifiable person. Pseudonymisation reduces risk but does not remove the data from privacy rules.
Who decides what counts as necessary?
The organisation collecting the data must be able to justify it against its stated purpose, and regulators can challenge that justification. For a specific situation, readers should consult the current law and regulator guidance, or a qualified lawyer.
How can a small team start?
List every field and log the product stores, write down why each exists, and delete or stop collecting anything without a clear answer. Then set a deletion schedule for what remains.
