Google Discover is the personalized feed in the Google app and on the Android home screen, and for many publishers and creators it is now a larger organic referral source than classic search — Google's last public figure, from 2018, put the feed at more than 800 million monthly users, and it has not published a newer number since. But it is not a surface you rank in. There is no query, no position, and no submission; Google assembles a feed from indexed content and matches cards to what it infers each person is interested in, and it says plainly that Discover traffic is less predictable than search. That makes a 'content strategy for Discover' a genuinely different discipline — you are building eligibility and fit for a recommender, not climbing a results page. The discipline got sharper on February 5, 2026, when Google shipped its first-ever Discover core update, explicitly tuned to reward three things: locally relevant content from sites in the user's own country, less sensationalism and clickbait, and in-depth, original, timely content from sites with real topic-by-topic expertise. At the same time the feed itself is becoming AI-mediated — reader-tunable by chat, increasingly fronted by AI summaries and topic overviews, and the subject of Google hiring to generate and grade AI content inside Discover. This guide is the practitioner's strategy for that surface: what Discover actually is and how it differs from search, what the February 2026 update rewards and penalizes, the eligibility and packaging signals a strategy has to satisfy, what the AI-mediated direction changes about what earns a click, and how to turn all of it into a repeatable editorial operation instead of a lottery ticket you buy one post at a time.
Google Discover is the personalized feed of cards that appears in the Google app and on the Android home screen, and for a lot of publishers and creators it has quietly become a bigger organic referral source than classic search — Google's last public figure, from 2018, put the feed at more than 800 million monthly users, and the company has not published a newer number since. That scale is why a Discover strategy is worth having. But it is also why the most common approach to it fails: people treat Discover like search, hunting for a keyword to target and a position to climb, and Discover has neither. There is no query, no ranking slot, and no submission. Google assembles the feed from indexed content and matches each card to what it infers a given person is interested in right now, and it says outright that Discover traffic is less predictable than keyword search.
So the real question is not 'how do I rank in Discover' — you cannot — but 'how do I build content Discover is eligible to surface, and give the feed enough fresh, on-topic candidates that being matched stops depending on luck.' That discipline got a sharp, official definition on February 5, 2026, when Google shipped its first-ever core update aimed specifically at Discover, tuned to reward locally relevant content, less clickbait, and genuine topic-by-topic depth. At the same time the surface is turning AI-mediated — reader-tunable by chat, increasingly fronted by AI summaries, and the subject of Google hiring to generate its own content for the feed. This guide is the strategy for that surface as it actually is in 2026: what rewards and what penalizes, the eligibility and packaging you control, what the AI shift changes about the click, and how to run it as an editorial operation rather than a lottery.
Start with the mental model, because importing the search model is where most Discover strategies go wrong. In Search there is a query a person typed, a ranking of results, and a position you try to occupy; you win by being the best answer to a known term. Discover inverts all three. Nobody searched. Google builds a personalized stream from indexed web content and creator media and decides, per person, which cards to show based on predicted interest — so the same article can be surfaced to one user and never shown to another with different interests. There is no term to target and no slot to win; a card is simply surfaced to someone, or it is not.
Google is unusually candid about the consequence. Its own documentation states that Discover traffic is less predictable and less dependable than keyword search, and that it fluctuates with shifting user interests, content-type adjustments, and algorithm updates. That is not a disclaimer to skim past — it is the defining property of the surface, and it has a direct strategic implication: you cannot engineer a guaranteed placement, so a strategy that stakes a quarter's traffic on one Discover-optimized post is fragile by construction. What you can engineer is eligibility and fit for a recommender, and volume of qualifying candidates. This is the same distribution-by-relevance logic that now governs the social feeds, worked through in social media discoverability beyond followers; Discover is Google's version of it, and the broader state of search in an AI world is the context it sits inside.
On February 5, 2026, Google announced its first core update built specifically for the systems that surface content in Discover, and the announcement is unusually concrete about intent. It names three goals, and because Google rarely spells out what a feed update is optimizing for, these three are the clearest strategic brief you will get for the surface. The rollout finished in late February after running longer than Google's original estimate, and it began with English-language users in the United States, with other countries and languages to follow — so if you are outside the US, the effects reach you later, and US traffic to content written for American audiences from a foreign site may shift before your home market does. The news-side write-up of the launch sits alongside the wider cross-surface changes in Google's personalization across Search, Discover, and News.
The update explicitly favors content from sites based in the user's own country, so that the feed surfaces more locally relevant stories. The strategic read is not 'add a city name to every headline'; it is that genuine local grounding — a domestic site covering subjects with real regional context — is now a positive signal where generic, placeless content is not. For a brand or creator with an actual geography, that is a reason to make the local angle explicit rather than scrubbing it out in a bid for a borderless audience. For a site writing for a foreign market, it is a warning that borrowed-audience traffic is exactly what this lever is tuned to reweight.
Google says the update reduces sensational content and clickbait — the headline that over-promises, manufactures outrage, or teases a payoff the article never delivers. This formalizes guidance Google has long given for Discover: titles and headlines should capture the essence of the content, and must avoid clickbait and sensationalism rather than manufacture a tap through exaggeration. The strategic consequence is that the headline-lottery approach — writing the most inflammatory title you can and hoping the feed rewards it — is now the pattern the update is built to suppress. Curiosity the content pays off is still rewarded; curiosity it betrays is a liability, and the gap between the two is exactly what the system is learning to score.
This is the lever that should reshape a strategy the most. The update surfaces more in-depth, original, timely content from sites with genuine expertise in a subject — and, critically, Google says expertise is judged topic by topic. Its own example: a general news site with a strong gardening section can qualify for gardening, while a movie-review site that published one gardening article likely will not. That is a direct instruction to own subjects rather than scatter across them. A site that covers one topic deeply and consistently builds the topic-level expertise the update rewards; a site that posts one piece on a subject and moves on does not accumulate the signal. Depth, originality, and consistency on a defined set of topics is the content strategy this update describes almost verbatim — and it rhymes with what every AI-mediated surface now rewards, argued in the AI content flood and signal quality.
Underneath the editorial direction sits a technical floor, and it is worth getting exactly right because it is the part you fully control. A page becomes eligible for Discover once it is indexed and complies with Discover's content policies — there is no special tag, sitemap, or markup that submits it. But eligibility is not appearance, and Google weights several packaging signals differently than search does. The first is the image. Every Discover card is an image card, so the hero image is a primary factor in whether a card surfaces at all, not a decoration. Google's published advice is specific: use a high-quality image at least 1200 pixels wide, enable the max-image-preview:large meta setting so Google can render a large preview, and prefer original, compelling imagery over generic stock or text-heavy graphics. A strong piece with a weak or missing large image routinely does not surface.
The second is freshness. Discover is organized around current interests, so it leans toward recently published content far more than evergreen search does — a timely, of-the-moment piece frequently outperforms an older, more authoritative one that no longer maps to what people care about today. This does not bar evergreen content, and refreshing a strong older piece can help, but it makes publishing cadence a genuine, strategy-level lever rather than a vanity metric: the more recent, on-topic candidates you supply, the more chances the feed has to match one to a reader. The third is page experience and E-E-A-T — fast mobile load, healthy Core Web Vitals, and content that reads as credible and clearly on a coherent subject, which is what lets Google classify your site and match it to the right interest cluster. And the fourth is reach: because Discover samples native video and social as well as article links, a strategy confined to one owned blog is invisible to large parts of the feed. The video-specific version of this entire eligibility checklist is in Google Discover video optimization.
A 2026 Discover strategy cannot stop at the February update, because the surface itself is changing shape underneath it in three ways at once, and all three pressure the single thing most strategies assume: the guaranteed click. First, readers can now steer the feed by conversation. Google began rolling out an AI personalization layer that lets a person tap a story's menu and tell it, in plain language, what topics and content types they want more or less of, then remembers the instruction. That turns Discover from one feed Google assembles into millions of reader-tuned micro-streams — and it moves the optimization unit from winning a single card to being the recognizable, on-topic source a reader deliberately asks the feed for more of. The full mechanics of that shift are in Google Discover's AI chatbot-tuned feed.
Second, AI summaries increasingly sit between the story and the click. Discover now groups publishers under short AI-written blurbs, and Google is testing a 'topic overviews' layer — a 'Dive deeper' button that opens an AI summary with links out to related coverage, which it began testing on video cards after a VP announced it in September 2026 (Google's Discover topic-overviews test). The strategic consequence is the same as it is for AI Overviews in search: if a two-sentence summary can convey the informational value, the click only survives for what the summary cannot reconstruct — a distinctive point of view, first-hand reporting or experience, a recognizable voice, or media the blurb can only reference. Generic, interchangeable coverage is precisely what a summary absorbs without a click, a dynamic traced in the publisher traffic collapse.
Third — and most speculatively — Google has posted engineering and design roles to generate and quality-check AI content inside Discover itself, turning a surface that has only ever recommended other people's pages into one that may also produce its own (Google is hiring to generate AI content inside Discover). No launch date, feature list, or share-of-feed figure is attached, and job descriptions are not roadmaps, so this is a direction to plan around, not a shipped product to react to. But the bar those roles describe is instructive: factuality, freshness, coherence, and tone checks on generated content. Thin, templated, obviously-synthetic content is exactly what such systems are built to filter out — which means the durable posture is to be the distinctive, clearly-attributed human source an AI-mediated feed elevates rather than replaces.
Put the February update and the AI shift together and the strategy resolves into one shape. Discover rewards a site that covers a defined set of topics with genuine depth, originality, and local grounding; packages every piece with a strong large image, an honest headline, and good page experience; publishes on a cadence that keeps fresh candidates in the feed; and is recognizable on those topics across the article, video, and social surfaces Discover samples — with the whole thing anchored to channels you own, because owned channels are where a reader a tuned feed routes to you can actually be kept. That is not a headline tactic. It is an editorial operation.
And it is precisely where most publishers and creators stall, because the operation is a volume-and-consistency demand that outruns hand production. Owning three or four topics deeply means shipping multiple on-topic, well-packaged pieces a week; being present across article, video, and social multiplies that by the number of surfaces; and the freshness lever means the pipeline can never go quiet for long without the feed losing its supply of candidates. The measurable failure mode is not a bad post — it is an empty pipeline, where you package one piece perfectly, it may or may not get picked up, and then nothing new arrives for weeks. A Discover strategy lives or dies on whether you can keep the operation fed.
Three habits actively work against a 2026 Discover strategy. Stop optimizing a single headline as if winning one card is the goal — the February update suppresses the sensational-title lottery, and a reader-steered feed increasingly decides what interests them regardless of your framing. Stop treating Discover as a guaranteed click channel; between reader tuning, AI summaries, and the feed's own unpredictability, banking a traffic plan on it is fragile, which is why owned channels have to be the durable layer. And stop spreading one thin post across many subjects: topic-by-topic expertise means a site that covers a subject once and leaves does not accumulate the signal the update rewards, while a site that owns the subject does.
The subtler trap is publishing competent, interchangeable content — the article that says exactly what every other article on the topic says. That is the content an AI summary absorbs without a click and the content a reader-tuned feed has no reason to route more of, and it is the pattern that stopped working across every AI-mediated surface, argued in why AI content stopped working. Specificity, a recognizable voice, local grounding, and first-hand substance are what survive both the summary and the tuning — and they happen to be exactly what the February update's 'original, in-depth, expert' language is reaching for.
Kompozy is an AI content generation and multi-platform publishing engine, and its fit for a Discover strategy is specific: the February update rewards topic-by-topic expertise produced with depth, originality, and consistency across the surfaces Discover samples — and that is an editorial-operation problem, not a one-post problem. Kompozy is built to run that operation. Point it at a topic you intend to own and it generates the full range the feed draws from: Blog Articles that give a subject a proper indexed home with surrounding depth, Email Newsletters for the list you own, and the native social and video pieces — Persona Shorts avatar video, Clipped Shorts cut from long-form, carousels, and quote or infographic images — that make you recognizable on that topic across the article, video, and social surfaces a tuned feed routes back to.
Two of Kompozy's constraints map directly onto what the update reweights. A single Persona Brief governs voice and a banned-word filter enforces it, so a week of output reads as one coherent, identity-bound source on a subject — the recognizability that both builds topic-level expertise and survives an AI summary that would otherwise absorb generic coverage. And because every card, hero image, and thumbnail is rendered through HyperFrames, the original, on-brand large imagery the image-card factor demands gets produced at the same volume as the copy, rather than defaulting to the generic stock that keeps a card from surfacing. The result is depth on a defined set of topics, packaged to Discover's spec, produced at a rate a hand-run operation cannot match.
Cadence — the freshness lever Discover weights so heavily — is Autopilot, which schedules and publishes the whole set across eight social platforms plus blog and email from one queue behind a per-post human review gate, so the feed is fed a steady supply of recent, on-topic candidates while you still approve everything that ships. Starter runs $199/mo and Pro $499/mo for creators and teams owning several topics; agencies running Discover strategies for clients use custom Enterprise. The boundary is worth stating plainly, because Discover invites overclaiming: Kompozy cannot submit anything to Discover, cannot guarantee the recommender surfaces a given piece, and does not control Google's indexing, the February update, or the AI-summary layer — no tool can, because there is no slot to buy and Google says eligibility is not appearance. What it removes is the production tax that makes a consistent, depth-led, well-packaged Discover operation impossible to sustain by hand, which is the specific reason most strategies die as an empty pipeline. The weekly mechanics of running that operation are in content distribution systems that compound.
A Google Discover content strategy in 2026 is a bet on being consistently, recognizably, and genuinely good on a defined set of topics — not on winning a single card. The February 2026 core update made the editorial brief explicit: local relevance, no clickbait, and topic-by-topic depth and originality. The technical floor is fixed and in your control: indexed pages, a strong 1200px image, honest headlines, good page experience, and a cadence that keeps fresh candidates in the feed. And the AI shift underneath it all — reader tuning, summaries in front of the click, the feed generating its own content — points to one durable conclusion: be the distinctive, owned, first-hand source an AI-mediated feed elevates, route the readers it sends you to channels you keep, and run it as an operation, because the strategy that loses is the one that optimizes one perfect post and lets the pipeline run dry.
You do not optimize a page to rank, because Discover has no query and no ranking position — it is a personalized feed Google matches to a user's inferred interests. You optimize for eligibility and fit: publish genuinely useful, original content on a consistent topic, meet the technical bar (indexed, good page experience, a large 1200px image enabled via max-image-preview:large), write headlines that spark interest without clickbait, and keep a regular cadence so the feed has fresh, on-topic candidates to surface. Google states eligibility does not guarantee appearance, so you are improving the odds, not buying a slot.
Announced February 5, 2026, it was Google's first-ever core update aimed specifically at Discover. It rewards three things: more locally relevant content from sites based in the user's own country, less sensational content and clickbait, and more in-depth, original, timely content from sites with genuine expertise in a subject. Expertise is judged topic by topic — a general news site with a strong gardening section can qualify for gardening, while a site with one off-topic article likely will not. It rolled out first to English-language US users, with other countries and languages to follow.
Yes, meaningfully, even though they share a technical foundation. SEO targets a query and a ranking position; Discover has neither. The feed matches a card to a person who did not search, so the weighting shifts: freshness matters far more than in evergreen search, the hero image is a first-class factor rather than a nicety, headlines must earn a tap mid-scroll rather than a click on a results page, and topical consistency — whether Google clearly understands what your site is about — decides whether you get matched to the right interest cluster. The eligibility work overlaps with SEO; the levers that decide whether a card surfaces are closer to a recommendation feed.
No, and any tool or agency that promises it is misleading you. There is no submission, no slot to buy, and no position to occupy. Google is explicit that meeting Discover's eligibility requirements does not guarantee an appearance and that Discover traffic is less predictable and less dependable than keyword search, with fluctuations driven by shifting user interests and algorithm updates. A strategy can maximize eligibility, fit, and the number of fresh, on-topic candidates you give the feed — it cannot force the recommender to pick any specific one.
Both, and increasingly beyond articles. Discover surfaces native video — especially YouTube and Shorts — and a growing share of short-form and social content alongside traditional article links, and Google has said it is expanding the kinds of creator content it shows. A strategy limited to one's own blog is invisible to the parts of the feed that sample video and social, so being recognizable on a topic across the formats and platforms Discover draws from gives a reader's interest more surfaces that can route back to you.
Three ways at once. Readers can now tune the feed by chat, telling Google in plain language what topics they want more or less of. AI summaries and a tested 'topic overviews' layer increasingly sit in front of cards, so a reader can absorb the gist without a click. And Google has posted roles to generate and quality-check AI content inside Discover itself. The strategic response is the same across all three: be a distinctive, clearly-attributed source whose point of view and first-hand substance a summary cannot replace, and anchor the relationship to channels you own.
Google Discover is a personalized feed, not a search ranking, so you earn visibility by being a consistent, recognizable source on a topic rather than by targeting a keyword. A winning 2026 strategy follows what the February 2026 core update rewards — locally relevant content, no clickbait, and genuine topic-by-topic expertise and depth — packaged with freshness, strong 1200px images, and honest headlines, and published on a cadence the feed can keep sampling. Google says eligibility never guarantees appearance, so a strategy improves the odds, not a slot.
Get started → · ← All guides · Compare Kompozy vs other tools