Does Google Discover Carry Sponsored Content?

What is Discover, and how does content reach it?

A feed rather than a search result. Discover surfaces content to people who did not search for it, based on inferred interests, which makes it the one Google surface where a reader encounters an article without having asked a question.

That difference matters for how the policies are written. Discover's content policies are almost entirely a list of what is not allowed: dangerous content, deceptive practices, harassing content, hateful content, manipulated media, medical content, sexually explicit content, violent extremist content, violent and gory content, and vulgar language.

What the documentation does not do is describe how content becomes eligible. It says content should follow best practices and must not violate the listed policies, and stops there.

This is the same shape found repeatedly in this series. Platforms publish disqualifying conditions and withhold selection criteria, which means nobody can sell a route in. A vendor promising Discover placement is promising an outcome with no documented mechanism behind it.

The feed format also changes the stakes of disclosure. A reader who searched for something arrived with intent and some context; a reader served an article in a feed has neither, and is correspondingly less equipped to notice that what looks like reporting was paid for. A disclosure rule matters more, not less, where the audience did not go looking.

What does it require of sponsored content?

Disclosure, in wording that will look familiar to anyone who has read the Google News policy examined elsewhere in this series.

Discover states that sponsorship “including, but not limited to, ownership or affiliate interest, payment, or material support, should be clearly disclosed to readers”. The definition reaches beyond cash to ownership stakes, affiliate arrangements and support in kind.

And the prohibition: “We do not allow content that conceals or misrepresents sponsored content as independent, editorial content.”

The near-identical phrasing across two separate Google surfaces is itself the finding. This is not a quirk of one product's rulebook; it is a settled position applied wherever Google distributes publisher content, which makes it considerably harder to treat as an edge case.

The repetition also removes a common negotiating position. A publisher pressed to run something without a label can no longer treat the requirement as one platform's idiosyncrasy, because the identical sentence governs the other surface it depends on for traffic.

What transparency does it demand of the publication?

Four specific things, each checkable in seconds on any outlet a buyer is considering.

News sources must provide clear dates and bylines; information about the authors, publication and publisher; and contact information. That is a short list and a surprisingly effective filter.

A great many sites sold in placement packages fail it. No named editor, no ownership disclosure, no contact route beyond a form, and articles carrying a generic staff byline or none at all. Those are not marginal deficiencies under this policy; they are the named requirements.

The convergence with journalism's own standards is again striking. The Trust Project's first two indicators ask who is behind the news and who made a given piece, and this policy asks for precisely that information in precisely that order. A platform and a professional standards body arriving independently at the same four items suggests the items are not arbitrary.

The contact-information requirement is the one most often missing and the most telling. A publication that cannot be reached is a publication that cannot be asked for a correction, cannot be queried about ownership, and cannot be held to anything it publishes — which is precisely why it appears on the list.

What about headlines and preview text?

There is a rule aimed squarely at the gap between a promise and a payload: Discover does not allow “preview content that misleads users to engage with it by promising details which aren't reflected in the underlying content”.

This is the clickbait provision, and it has a specific relevance to sponsored material. A placement's headline is frequently the most heavily optimised element, written to earn a click in a feed, and the article beneath it is often thinner than the headline implies because its purpose was the placement rather than the reporting.

The rule is about the relationship between the two rather than about sensationalism as such. A dramatic headline supported by the article is fine; a modest headline promising specifics the piece never delivers is not.

For a buyer this is a useful drafting test. If the headline would disappoint a reader who read to the end, the piece has a problem that no distribution can fix, and on this surface it is a stated policy breach rather than a matter of taste.

The rule also bears on how placements are reported back to clients. A headline written to perform in a feed is often the metric a campaign is judged by, and optimising it away from what the article supports is the exact behaviour the policy names.

Why does one surface's policy matter to a placement buyer?

Because Discover is a meaningful traffic source for publishers, and a placement inherits its host's standing on it.

A publication excluded from Discover, or performing poorly there, delivers less than a buyer expects from a site with comparable search visibility. And exclusion decisions are made on the basis of the publication's whole behaviour, including how it handles sponsored content generally — not only the piece a given buyer paid for.

That creates an exposure worth naming. A buyer placing coverage on an outlet that routinely runs undisclosed sponsored material is buying into a domain carrying policy risk it did not create and cannot control.

It also means the outlets most attractive on price are often the ones least likely to satisfy these requirements, since the absence of named editors, ownership disclosure and consistent labelling is part of what makes cheap placement cheap.

There is also a portfolio effect worth weighing. Buyers frequently purchase several placements at once across a tier of similar outlets, which means a single package can concentrate exposure in exactly the kind of site these requirements exclude, rather than spreading it.

It is also the surface where a buyer is least able to verify what happened. Search results can be checked by searching; a feed cannot be audited from outside, so any claim about Discover performance rests entirely on what the publisher reports.

What this means for anyone buying placement

Run the four-item transparency check before buying: dates, bylines, information about authors and publisher, and contact details. It takes a minute and it is the platform's own list.

Expect disclosure and treat it as the product working. The same requirement now appears in Google News, in Discover, in FTC guidance and in journalism's own standards, which means the undisclosed version is not a cheaper variant of the same thing but a different and riskier one.

Write headlines the article can support. The preview rule is stated, and a headline promising what the piece does not deliver is a policy breach as well as a poor reader experience.

And discount promises of Discover placement entirely. The policies describe what disqualifies content and never describe what qualifies it, so there is nothing for a vendor to execute. Media Morpheus sells placement, and what we can honestly point to is whether an outlet meets the published requirements — which is checkable, unlike everything downstream of it.

None of this makes Discover a bad outcome to hope for. It makes it an outcome rather than a deliverable, which is the distinction this series keeps returning to.

Methodology and limitations

This report is based on Google's published Discover content policies, with comparison to Google News policies in Publisher Center Help, quoted for the sponsorship disclosure requirement, the concealment prohibition, transparency requirements and the preview content rule. Four limits apply. First, the policies state what disqualifies content and do not describe eligibility, so nothing here explains why any article appears in Discover. Second, the policy governs publications; enforcement consequences fall on the outlet rather than on an individual advertiser, though a buyer's coverage is affected by its host's standing. Third, no data is reported on Discover traffic, exclusion rates or enforcement, because Google publishes none and no reliable independent figure was located. Fourth, platform policies change without notice and this describes the position at the date of publication.

Conclusion

Discover carries sponsored content and refuses to carry sponsored content pretending to be editorial, in language almost identical to the Google News policy. The repetition is the most useful thing about it: two separate surfaces, one settled requirement, which makes disclosure a condition of Google distributing publisher content rather than a rule of one product. Alongside it sits a four-item transparency list — dates, bylines, information about authors and publisher, contact details — that quietly disqualifies a large share of the outlets sold cheaply in placement packages, and a preview rule requiring a headline to be supported by the article beneath it. As everywhere else in this series, the policies say clearly what will exclude you and say nothing at all about how to get in.

Frequently Asked Questions

Sources

1. official Google. Discover content policies, for policy categories, sponsorship disclosure, transparency requirements and the preview content rule.
2. official Google. Google News policies, cited for the parallel sponsorship language.
3. trade body The Trust Project. Trust Indicators, cited for the overlapping transparency expectations.