What Is Data Minimisation? – Definition and Practice
Learn what data minimisation requires, how it applies to collected web data, and the practices that reduce risk without reducing usefulness.
Data minimisation is collecting and keeping only the data a defined purpose requires. It is a privacy principle codified in several regulatory frameworks, and it is also straightforwardly good engineering.
1. The Principle
Data should be adequate, relevant, and limited to what is necessary for the stated purpose. Three obligations follow:
- Define the purpose first, before deciding what to collect.
- Collect only what serves it.
- Keep it only as long as it serves it.
"We might need it later" is precisely the reasoning the principle exists to exclude.
2. Why It Applies Especially to Collected Data
Collection is cheap and indiscriminate. A page fetch yields everything on the page, and storing all of it is easier than deciding what matters.
That default accumulates personal data nobody planned to hold — names, contact details, review authorship — with the obligations that come with it and none of the intent.
3. Practices That Minimise Without Losing Value
| Practice | Effect |
|---|---|
| Extract defined fields | Store what the purpose needs, not the whole page |
| Aggregate early | Keep counts and distributions rather than individuals |
| Truncate or hash identifiers | Retain the ability to deduplicate without holding the value |
| Separate raw from derived | Retain raw briefly, derived longer |
| Set retention per field | A price needs history; a reviewer's name usually does not |
4. Raw Retention Versus Minimisation
There is a genuine tension. Keeping raw payloads makes extraction improvable, which matters because web sources cannot be re-read. Minimisation argues for keeping less.
The workable resolution is a short retention window for raw payloads containing personal data, longer retention for derived non-personal records, and documented reasons for both.
5. Personal Data Deserves Specific Care
Public availability is not consent, and jurisdictions differ on what may be collected and used.
- Collect the minimum that serves the purpose.
- Keep it separable, so it can be deleted without destroying the rest.
- Secure it as you would any personal data.
- Honour deletion requests, which requires knowing where it is.
- Do not use it for a purpose it was not collected for.
6. Retention Requires Deletion That Works
A retention policy nobody enforces is not a policy. Deletion must be implemented — scheduled, verified, and covering backups and derived copies.
The common failure is a documented ninety-day window with data from three years ago still present, which is worse than having no policy because it misrepresents the position.
7. It Reduces More Than Legal Risk
Less data is cheaper to store, faster to query, simpler to secure, and less damaging if exposed. Minimisation is one of the few privacy practices whose engineering benefits are immediate.
That makes it easier to justify than most compliance work, and easier to sustain.
8. Summary
Data minimisation means defining a purpose, collecting only what it requires, and keeping it only as long as it is needed.
For collected web data it means extracting defined fields rather than storing whole pages, keeping personal data separable with short retention, and implementing deletion that demonstrably works — with real engineering benefits alongside the compliance ones.