Consent, Contract, or Legitimate Interest: Three Privacy Law Bases Compared
Archi4 min read
Contents
The headline says fine. The filing says legal basis.#
Another week, another privacy enforcement story. The coverage usually leads with the number: a few million in penalties, a regulator's quote, a promise to appeal. What gets skimmed past is the part that decides the case: which legal basis the company claimed for processing personal data, and whether it could defend that choice in writing.
Almost every consent banner, cookie notice and privacy policy in Europe and the UK rests on one of three options under the General Data Protection Regulation: consent, contractual necessity, or legitimate interest. They are not interchangeable. Each has a different evidentiary burden, a different shelf life, and a different exit cost when something changes.
Here is how the three compare, and why the distinction matters more than the fine totals.
Option one: consent#
Consent is the most visible basis and the most fragile. It requires a clear affirmative act, a specific purpose, and a way to withdraw that is as easy as the way to give it. A pre-ticked box does not count. A banner that says continuing to browse means acceptance does not count either, no matter how many sites still use it.
The hidden cost is operational. Every withdrawal is a processing event that must be honored, logged and propagated to vendors. If a company collects consent in 2021 and changes its data flows in 2024, the original consent may no longer cover the new use. Re-consenting a user base of several million is slow and expensive, and the drop-off rate is real.
Consent works best for optional processing: marketing emails, analytics that go beyond what is strictly needed, personalization features. It is a poor fit for anything the business cannot operate without, because a user who says no has to be served something else entirely.
Where consent cases actually fail#
Regulators rarely object to consent in principle. They object to records. In a typical enforcement, the company cannot show who consented, when, to what wording, or on which version of the form. The banner looked fine. The audit trail did not exist.
Option two: contractual necessity#
This basis covers processing needed to perform a contract with the person, or to take steps at their request before a contract. Shipping an order to an address is the clean example. The address is necessary. The customer cannot receive the parcel without giving it.
The temptation is to stretch the category. Companies have tried to classify ad targeting as contract performance on the theory that the service is free because advertising pays for it. Regulators have pushed back repeatedly. Necessity means the contract cannot be performed without the data, not that the business model depends on it.
The advantage is durability. There is no withdrawal right, no expiry date, and no re-collection cycle. The disadvantage is that the scope is narrow and documented externally, in the contract itself. If the service changes, the contract language has to change with it, and customers notice when the terms shift.
Option three: legitimate interest#
Legitimate interest is the most flexible and the most litigated. It requires a three-part test: a genuine interest, necessity of the processing, and a balancing exercise against the individual's rights and expectations. Fraud prevention, network security and some internal administration routinely pass. Behavioral advertising rarely does without strong safeguards.
The practical burden falls on documentation. A legitimate interest assessment is a written record that has to be produced on request, and it has to be specific to the processing, not a template copied from a consultancy deck. Companies that win on this basis usually have assessments that name the data, the retention period, the safeguards and the reasoning.
A useful property: legitimate interest survives business change better than consent. It does not require re-consenting a user base. But it is not a fallback for when consent fails. Switching bases midstream, after a consent campaign underperforms, requires a fresh assessment and often a fresh notice to users.
What the coverage keeps missing#
The fine is the least interesting number in an enforcement decision. The more revealing detail is the switch. Watch for cases where a company moved from consent to legitimate interest after sign-ups fell, or from legitimate interest to consent after a complaint. Each switch leaves a paper trail, and regulators read it closely.
The second detail is retention. A basis that justifies collection does not automatically justify keeping the data. Many enforcement actions turn on records held beyond any stated period, with no deletion job running.
The third is vendor flow. A company's own basis does not automatically cover its processors. Each onward transfer needs its own lawful footing, and the contracts have to say so.
Choosing between them#
The short version: use consent for optional things, contract for things the customer is buying, and legitimate interest for security and fraud work with a written assessment behind it. Do not treat any of the three as a default that covers the gaps.
If a company cannot explain in one paragraph why it picked a basis, and produce the record that shows the choice was made deliberately, the fine is not the risk. The risk is that the record has to be built years later, from memory, while a regulator waits.
More articles