Google Discover SEO is the practice of making your content eligible for Google’s personalised mobile content feed, which surfaces articles to users based on their interests rather than in response to a search query. Unlike search SEO, where you optimise a page against a known query and a measurable position, Discover has no query and no ranking — you either meet the eligibility conditions or you are invisible.
Google’s documentation states that content “is automatically eligible to appear in Discover if it is indexed by Google and meets Discover’s content policies.” There is no submission process and no Google News membership requirement — a misconception that keeps most non-publisher sites from ever looking at the channel.
Google recommends Discover images be at least 1,200 pixels wide, with more than 300,000 total pixels and a 16:9 aspect ratio, and enabled via the max-image-preview:large robots directive. In February 2026, Google shipped a Discover-specific core update — the first of its kind — targeting clickbait and sensationalism while favouring original, in-depth, timely content from sites with demonstrated expertise in a topic.
The reason this channel gets ignored by everyone outside publishing is that all the literature about it was written by publishers. Newsroom advice about story velocity and breaking coverage doesn’t translate to a B2B SaaS site, so marketing teams read one article and conclude Discover isn’t for them.
That conclusion is usually wrong. But so is the opposite reaction — treating Discover as a growth channel you can plan around. Let me explain what it actually is, and how to decide whether it deserves your attention.
What Google Discover actually is
Discover is a query-less feed. It appears in the Google app on iOS and Android and on the home screens of most Android devices, and it shows content Google predicts a specific user wants to see, based on that user’s activity, follows and interests.
That single structural difference changes everything downstream. In search, demand exists before your content does — someone types a query, and you compete for a position against a known set of competitors. In Discover, there is no demand signal to target. Google decides, per user, whether your article is interesting enough to inject into their feed.
This has three practical consequences that most guides skip:
There is no position to defend. You can’t be “ranking third” in Discover. A page is either being served to some users or it isn’t, and the population it’s being served to changes continuously.
Traffic arrives as spikes, not curves. A piece that gets picked up can deliver more sessions in 48 hours than the page will earn from search in a year, and then flatten to zero. That is normal behaviour, not a penalty.
It is overwhelmingly mobile. Discover lives in mobile surfaces. If your site’s mobile experience is poor, the channel is closed to you regardless of how good the content is.
In my experience auditing sites that suddenly gained or lost Discover traffic, the team’s first instinct is almost always to look for a ranking change. There usually isn’t one. The search performance is flat and the Discover line moved on its own, because the two systems make independent decisions.
What the February 2026 Discover core update changed
Google announced a Discover core update in February 2026 — significant because it was the first core update aimed specifically at Discover rather than at search results generally. It confirmed something practitioners had suspected for a while: Discover is now ranked by its own system, tuned for interest-based consumption rather than query satisfaction.
The stated direction of the update was to reduce clickbait and sensationalised content, surface more locally relevant material, and favour original, in-depth and timely content from sites with real expertise in the topic they’re covering.
The consequence that matters most for brands is the emphasis on originality. Content whose primary function is to summarise what already exists elsewhere lost ground. That describes a very large share of the “ultimate guide” content marketing teams produce — pages assembled from the top ten results for a keyword, with nothing in them that didn’t already exist on the web.
If you want Discover visibility, the content has to contain something that isn’t available anywhere else: your own data, your own testing, a decision framework from actual client work, or first-hand experience of the thing you’re describing. This is the same pressure E-E-A-T applies in search, applied harder, and it is why generic AI-assembled content performs so badly in this channel.
The Google Discover SEO eligibility checklist
Discover eligibility is mechanical. These are the conditions to satisfy before anything else matters — none of them guarantee traffic, but failing any one of them guarantees none.
- Confirm the page is indexed. Discover eligibility starts with Google’s index. If the URL isn’t indexed, nothing downstream is relevant. Check it in the URL Inspection tool, not by a
site:search. - Serve a featured image of at least 1,200 px wide. Target more than 300,000 total pixels and a 16:9 aspect ratio. A site logo or a generic stock graphic in the metadata will not qualify — Google’s policies call this out explicitly.
- Enable
max-image-preview:large. Add it to the robots meta tag or theX-Robots-Tagheader. Without it, Google will not render your image at the size Discover requires, which effectively removes the page from consideration. This is the single most common failure I find, and it is usually a template regression nobody noticed. - Pass Core Web Vitals on mobile field data. LCP under 2.5 seconds, INP under 200 milliseconds, CLS under 0.1. Use the field data in Search Console, not a lab score — Core Web Vitals are assessed on what real users experience.
- Attribute the content to a real, identifiable author. A named byline with a genuine author page and verifiable credentials. Anonymous content struggles in Discover in a way it doesn’t always struggle in search.
- Add
Articlestructured data. It is not a documented requirement, but it removes ambiguity about what the page is, when it was published, and who wrote it. Get your schema markup right and you stop relying on Google’s inference. - Write headlines that describe rather than tease. The February 2026 update explicitly targeted clickbait. A headline that overstates what the page delivers is now a liability, not a CTR tactic.
Run this as a checklist over your top twenty pages before you write anything new. In most audits I run, points 3 and 4 are already broken sitewide, and fixing them costs a developer an afternoon.
Why Discover traffic disappears without a ranking drop
This is the scenario that generates panic. Traffic halves overnight, Search Console shows no ranking movement, no manual action, no crawl errors — and no obvious cause.
Almost always it’s one of four things.
A story simply aged out. Discover favours freshness. A piece that was surfacing to hundreds of thousands of users stops being interesting, and the feed moves on. Nothing broke.
A template change stripped the eligibility signals. A CMS update that drops max-image-preview:large, swaps the featured image for something under 1,200 px, or removes the author byline will cut Discover off without touching search rankings at all. Search tolerates these changes; Discover doesn’t.
Mobile performance regressed. A new script, a heavier hero image, a third-party tag added by the marketing team. Search rankings absorb a small Core Web Vitals regression. Discover is less forgiving because the entire surface is mobile.
An algorithmic update hit. The February 2026 core update caused exactly this pattern for sites leaning on summarised or sensationalised content. If your drop aligns with a known update date, you’re looking at a content problem, not a technical one, and the fix is measured in quarters.
The diagnostic order matters: check the update calendar first, then diff your page templates, then check field-data Core Web Vitals, and only then conclude the content aged out. Working the other way round wastes weeks. A structured technical SEO review will surface template regressions faster than any amount of content analysis.
How to measure Google Discover SEO performance honestly
Discover reporting is deliberately limited, and the limitations tell you something about the channel.
A Discover tab appears in Google Search Console automatically once your site crosses a traffic threshold. It reports impressions, clicks, and CTR by page and date. It reports no queries, because there are none.
That absence is the whole point. You cannot do keyword research for Discover. You cannot forecast it. You cannot build a content plan that targets Discover volume, because the volume is a function of how many interested users Google decides to show your page to — a number you have no visibility into and no influence over beyond making the content genuinely worth surfacing.
So measure it as a bonus channel, not a planned one:
- Track Discover clicks as a separate line, never blended into “organic”. Blending them hides both a Discover spike and a search decline.
- Segment Discover sessions in GA4 so you can see whether that traffic converts. Usually it converts far worse than search traffic, because the user had no intent — they were scrolling.
- Judge Discover by the eligibility rate of your published pages, not by session count. “What percentage of our new pages meet all seven conditions?” is a controllable metric. “How many Discover sessions did we get?” is not.
That last distinction is the one I push hardest with clients. Optimise the inputs you control and treat the output as weather.
Should your site invest in Discover at all?
Here’s the honest consultant answer, and it’s not “yes, absolutely, here’s a retainer.”
Do the eligibility work regardless. Points 1 through 7 above are cheap, one-time, and mostly benefit search anyway. Correct image sizing, max-image-preview:large, passing Core Web Vitals and real author attribution are things you should be doing whether Discover exists or not. There is no scenario where this is wasted effort.
Do not build a strategy on Discover volume. I have watched teams see one Discover spike, extrapolate it into an annual forecast, and restructure a content calendar around a channel that can go to zero in a week without warning. Discover is high reward and high variance, and variance is the part that gets skipped in the pitch deck.
Reconsider if you publish timely, interest-driven content. If your business genuinely produces original, topical material — original research, industry analysis, product category expertise — then Discover is a real distribution surface worth deliberate attention, and it pairs naturally with a serious content SEO programme.
Ignore the volume entirely if you publish evergreen commercial pages. A B2B site with twelve service pages and a quarterly blog post is not going to build a Discover habit. Do the eligibility hygiene and spend the strategic effort elsewhere.
The framing I’d leave you with is this: Google Discover SEO is an eligibility discipline, not a ranking discipline. You engineer the conditions under which Google can choose your content, and then you accept that the choosing isn’t yours. Teams that internalise that get the upside without building a plan on sand — and teams that don’t end up explaining a traffic cliff that no ranking report can account for.
If you want a second opinion on whether your site is technically eligible in the first place, that’s the kind of thing an SEO audit answers in a week.
Frequently Asked Questions
How do I get my content on Google Discover?
You don’t submit to Google Discover — content becomes eligible automatically once it is indexed by Google and complies with the Discover content policies. Beyond that baseline, the practical requirements are a featured image at least 1,200 pixels wide with the max-image-preview:large robots directive enabled, fast mobile performance, and a clearly attributed author. There is no application, no approval queue, and no Google News membership required.
Why did my Google Discover traffic drop?
Discover traffic usually drops for one of three reasons: an algorithmic change, a technical regression on mobile, or simply the natural decay of an individual story’s relevance. Google rolled out a Discover-specific core update in February 2026 that demoted content which mostly summarises what already exists, so check whether your drop lines up with that date. If it doesn’t, verify your Core Web Vitals on mobile and confirm max-image-preview:large is still being served — a template change that strips it will kill Discover eligibility silently.
Does Google Discover only work for news sites?
No. Google’s documentation says content is eligible if it is indexed and meets the content policies — there is no requirement to be a news publisher or to be accepted into Google News. In practice most Discover coverage is written by and for publishers, which creates the impression it’s a news-only channel. Evergreen guides, product explainers, and interest-driven brand content all appear in Discover regularly.
How do I see Google Discover traffic in Search Console?
A Discover tab appears automatically in the Search Console Performance section once your site has received enough Discover traffic to report. If you don’t see the tab, your site has not met the threshold — it is not a settings problem. The report shows impressions, clicks and CTR by page and by date, but no query data, because Discover has no queries behind it.
How long does it take to appear in Google Discover?
There is no fixed timeline, and many sites never appear at all. Discover surfaces content while it is fresh, so a piece typically either gets picked up within hours to days of being indexed, or not at all. Building consistent Discover visibility is a matter of publishing regularly in a defined topic area over months, not of optimising a single page.
