Google Discover shows your content to people who never searched for it. There is no query, no keyword and no search box involved: Discover is a feed of articles and videos matched to a person’s interests, drawn from their Web and App Activity.
For publishers it is one of the few remaining ways to reach a large audience without ranking for a keyword, and it is also one of the least predictable. This guide explains how Discover works, what Google actually requires, and the checklist that gives a page its best shot.
The first thing to understand is that Discover is part of Google Search, not a separate product with its own submission form. There is nothing to apply for and no tag to add. If your content is indexed by Google and meets Discover’s content policies, it is automatically eligible. Eligible, however, is not a promise: Google is explicit that eligibility is not a guarantee of appearing.
Discover Is Not Search: No Query, Only Interests
Traditional search starts with a question. The user types something, and Google ranks pages against it. Discover starts with a person. Google builds a picture of what someone is interested in from their Web and App Activity, including their search history, and then shows content related to those interests in a feed.
That difference changes everything about how a page gets found. In search, you can target a query and measure a ranking. In Discover, there is no query to target and no ranking to track. A page appears because it matches someone’s interests at the moment they open the feed, which is why two people with different histories see entirely different Discover feeds.
Discover surfaces appear in the Google app and in mobile Chrome, and the feed is a mobile experience first. That is part of why images matter so much here: on a phone, the image is often what a person notices before the headline, and Google’s own guidance treats large, high-quality images as one of the strongest levers a publisher controls.
The Two Conditions for Eligibility
Google states the eligibility rule in one sentence: content is automatically eligible to appear in Discover if it is indexed by Google and meets Discover’s content policies. No special tags or structured data are required.
The first condition, indexing, is the one most publishers already satisfy without thinking about it. If Google can crawl and index the page, the door is open. If a page is blocked from indexing, it cannot appear in Discover at all, whatever its quality. Indexing is the floor, not the achievement.
The second condition, the content policies, is where eligibility is actually won or lost. The policies cover advertising and sponsored content, misleading preview content, and transparency.
Sponsored material must not exceed the real content, and sponsorship must be clearly disclosed rather than dressed up as independent editorial. Preview content must not promise details the article does not deliver. And news sources should show clear dates and bylines, information about the authors, publication and publisher, and contact information.
Violations have consequences Google states plainly. Content that breaks the policies may be removed from Discover, and repeated or egregious violations can end a site’s eligibility entirely, including for surfaces such as “Articles for you” in Chrome. Manual actions for Discover violations appear in Search Console under Security and Manual actions, so a sudden disappearance has a place to be checked.
What Google Actually Recommends
Because Discover is part of Search, it uses many of the same signals and systems Search uses to judge helpful, people-first content. Google’s advice to publishers chasing Discover is, in essence, its general advice on creating helpful, reliable, people-first content, applied to a feed.
On top of that, Google recommends content that is timely for current interests, tells a story well, or provides unique insights. Timeliness matters because the feed is built around what people care about right now.
But it is not the only route in: Google notes that older content may appear if it is helpful and relevant to someone based on their interests. An evergreen explainer can surface months after publication when a person’s interests swing toward its topic.
The flip side is that Discover adjusts the types of content it shows over time to better match what people are looking for. Sports, health, entertainment and lifestyle content appear regularly, but the mix shifts, and a site whose traffic depends on one content type can feel those shifts sharply.
Images: The Biggest Lever You Control
If there is one technical recommendation to take from Google’s Discover documentation, it is the image specification. Google recommends compelling, high-quality images that are relevant to the content, especially large images, which it says are more likely to generate visits from Discover.
The specification has four parts. Images should be at least 1200 pixels wide, with a resolution of more than 300,000 total pixels, in a 16:9 aspect ratio, and enabled by the max-image-preview:large setting, or by using AMP. Google’s own example makes the arithmetic clear: a 16:9 image at 1280 by 720 pixels has 921,600 total pixels and satisfies the specification.
Two details are easy to get wrong. First, the ratio is a recommendation, not a hard bar: an image can clear the size and pixel thresholds in another ratio, but Google’s guidance favours 16:9 for Discover presentation.
Second, the image Google picks as the thumbnail can be influenced by specifying it in schema.org markup or the og:image meta tag, and Google warns against using generic images such as a site logo, or text-heavy images, in those fields. The thumbnail should be relevant and representative of the page.
For a publisher, the practical rule is simple: every article that could plausibly reach Discover deserves a real, large, relevant hero image, marked up properly, with the large-preview setting enabled. A page with a small or missing image is not disqualified, but it competes with one hand tied behind its back.
Headlines That Help, Not Trick
Headlines get their own section in Google’s guidance, and the message is consistent: capture the essence of the content. Google warns against clickbait and similar tactics that artificially inflate engagement, whether by misleading or exaggerated details in the title, snippet or image, or by withholding crucial information required to understand what the content is about.
Sensationalism gets the same treatment. Google advises against tactics that manipulate appeal by catering to morbid curiosity, titillation or outrage. A headline that promises more than the article delivers may win a click and lose the reader, and Google treats that pattern as a reason not to show the content again.
The honest version of a curiosity gap still works: a headline can be specific, surprising and specific to the story without deceiving anyone. The test is whether a reader who clicks feels the article delivered what the headline promised. If the answer is yes, the headline is doing its job for Discover as well as for the reader.
Remember that Google judges preview content as a whole: the title, the snippet and the image together. A truthful headline paired with a misleading thumbnail fails the same test as a misleading headline, and a snippet that promises a detail the article never delivers counts as misleading even if the title is clean.
Why Discover Traffic Rises and Falls
Google is unusually candid about Discover traffic: it is less predictable and less dependable than keyword-driven search visits, and publishers should treat it as supplemental to search traffic, not a replacement for it. The feed’s serendipitous nature means a page can surge one week and vanish the next without anything on the page changing.
Google names three reasons traffic changes. The first is changing interests: if people stop searching for a topic, their feeds show other content they are more interested in, and publishers covering the old topic lose traffic.
The second is content types: as Discover adjusts what kinds of content it shows, sites built around a falling type feel it. The third is updates to Google Search itself: because Discover is an extension of Search, core updates can move Discover traffic too.
The practical consequence is a planning rule. Discover traffic should be treated as a bonus on top of a search foundation, not as the foundation itself. A site that depends on Discover for its core audience is exposed to shifts it cannot control and cannot predict.
Measuring Discover in Search Console
Discover performance is measured in Search Console’s Performance report for Discover, which shows impressions, clicks and click-through rate for content that has appeared in Discover over the last 16 months, provided the data reaches a minimum threshold of impressions. The report fully tracks a site’s Discover traffic across all surfaces where users interact with Discover, and it includes traffic from Chrome.
That report is the only reliable way to know whether Discover is working for a site, and it is worth checking separately from search performance. A page can rank nowhere for any query and still collect Discover impressions, and a page can rank well while never appearing in a feed.
One caveat: the report only shows data once it reaches a minimum threshold of impressions. A small site may be eligible and even appearing, yet see nothing in the report until its Discover volume crosses that line. Absence from the report is not proof of absence from the feed.
Discover, Search and News Are Different Channels
It helps to keep the three channels distinct. Search answers a query the user typed. Discover shows content matched to interests without any query. News surfaces current, topical stories. A single article can appear in all three, but for different reasons and with different audiences.
The clearest difference is age. Search and News both favour freshness for topical queries, while Discover explicitly allows older content to appear when it matches someone’s interests. That is why an evergreen guide can have a long second life in Discover while its search traffic settles.
The measurement differs too. Search performance is tracked by query, Discover by page and feed impression, with no query attached. Treating the two reports as one channel hides what each is doing, and the fixes for a weak channel differ: search problems are usually about relevance and ranking, Discover problems are usually about images, headlines and policy standing.
For a fuller picture of how Search Console reports on newer surfaces, see our guide to Search Console’s generative AI reports, which covers the equivalent reporting for AI features in Search.
What Discover Filters Out
Discover is an interest feed, not an index dump, and Google filters what it will present. The feed seeks content suitable for interest-based browsing, such as articles and videos, and filters out content that might confuse readers or that is not wanted in a feed: job applications, petitions, forms, code repositories, and satirical content without context are Google’s own examples.
Beyond that, Discover uses SafeSearch and filters out content that might be seen as shocking or unexpected. The feed is designed to be opened casually on a phone, and content that jars in that context is less likely to be shown, whatever its quality.
There is also a separate route into feeds: following. If users choose to follow a website, its content appears in their Following feed, unless it violates Google Search’s overall policies or the policies for Search features.
Those policies cover dangerous, harassing, hateful, manipulated-media, medical, sexually explicit, terrorist, violent and vulgar content. A loyal following is the closest thing Discover offers to a guaranteed audience, and it is earned the same way as everything else: by publishing content people want to keep seeing.
A Practical Checklist for Publishers
First, make sure the page is indexed and the site is in good standing: no policy violations, no manual actions, and transparency basics in place, including dates, bylines, author and publisher information, and contact details.
Second, give every article a large, relevant hero image: at least 1200 pixels wide, more than 300,000 pixels in total, ideally 16:9, enabled with max-image-preview:large, and specified in schema.org markup or og:image. Avoid logos and text-heavy images as the chosen thumbnail.
Third, write headlines that capture the essence of the content, without clickbait, exaggeration, withheld essentials or sensationalism. Specific and honest beats vague and loud, in the feed as everywhere else.
Fourth, publish content that is timely, tells a story well, or offers a unique insight, and keep older content accurate, because a helpful older article can still surface when someone’s interests turn toward it.
Fifth, deliver a good page experience overall. Discover is a mobile feed, and a page that loads slowly or reads poorly on a phone will not convert the visit it earns, however good the thumbnail looks.
Sixth, keep an eye on Search Console for Discover manual actions, and fix any policy issue promptly. A violation that goes unaddressed can escalate from a single removed page to a site losing Discover eligibility, and the manual-actions report is where Google tells you what it found.
What Discover Cannot Do for You
Discover cannot be bought, submitted to, or guaranteed. There is no form, no tag and no structured data that puts a page in the feed, and no amount of optimisation promises an appearance. Eligibility is automatic for indexed, policy-compliant content, and appearance is Google’s decision, driven by people’s interests.
Discover also cannot replace search traffic. Its traffic is supplemental by design: serendipitous, interest-driven and less predictable than keyword visits. A publisher that builds its audience on search and treats Discover as a bonus gets the best of both. A publisher that builds on Discover alone is building on weather.
Used that way, Discover is genuinely valuable. It reaches people who would never have searched for your topic, at the moment they are open to it, with a large image and an honest headline doing the persuading. The work is not gaming the feed; it is publishing content worth recommending, presented the way the feed rewards.
Background
Discover’s rules have stayed notably stable: index the content, keep it within the policies, present it with a large image and an honest headline, and let people’s interests do the rest. The feed itself keeps changing around those rules, as interests shift and Search updates roll through, which is why the traffic it sends will always be less predictable than search. You can browse the subjects we cover on our topics index.
Sources: Google Search Central: “Get on Discover”; Google Search Help: “Discover content policies”.
0 Comments