What Google actually looks for

Google’s Right to Be Forgotten process is often described as opaque. In one sense it is: the decision-making is discretionary, and the reasoning in refusal letters can be thin. But the substantive criteria are published and consistent. Google assesses whether the information in a name-based search result is inaccurate, inadequate, irrelevant or excessive, and then weighs that against the public interest in continued access.

Each word does specific work

Inaccurate is narrow. It means factually wrong — not merely unflattering. A request built on the argument that an article is “unfair” will usually fail; one that demonstrates a concrete factual error has a basis.

Inadequate captures information that is technically true but incomplete in a way that misleads. A report of a charge without the later acquittal is the clearest example. So is an article that describes an arrest without recording that no charge followed.

Irrelevant is temporal and contextual. The information may have been relevant at publication but no longer is. This is where most successful requests live: the facts are accurate, but their continued prominence in a name search no longer serves any legitimate purpose.

Excessive is about proportionality. The data may be relevant, but its continued prominence in search results is not. An article that names an individual in a tangential paragraph, yet surfaces first for their name decades later, is a proportionality problem rather than an accuracy one.

Assertion versus demonstration

A request that says “this article is old” is weaker than one that explains why the passage of time has severed the link between the information and any legitimate public interest.

The strongest submissions do not simply assert these qualities; they demonstrate them with evidence and argument. That difference — between assertion and demonstration — is expertise. It is also, in practice, the difference between a refusal and a delisting.