Google Discover: how the feed works and how to get in
TL;DR
- Discover is a personalized feed with no search query - Google decides what to show each user by interest, freshness and quality.
- Eligibility, not ranking: you can't target a keyword, only become the kind of content Discover surfaces.
- Large, high-quality images (over 1200px) with max-image-preview:large are close to mandatory.
- Traffic is high-volume but volatile - build for consistency, not a single viral hit.
Google Discover is the feed of cards you see in the Google app and on Android - content Google surfaces based on your interests, with no search query at all. For publishers and content brands it can be enormous traffic. It's also the most volatile channel in search, precisely because you don't control the query - there isn't one.
Eligibility, not ranking
This is the mental shift. In search you optimize for a query and compete for a position. In Discover there's no query: Google predicts what each user wants and serves it. You can't rank for "X" - you can only become the kind of content Discover trusts enough to show. That means quality and interest signals, not keyword targeting.
What makes you eligible
- Large, high-quality images - over 1200px wide - enabled with <code>max-image-preview:large</code>. Small images alone can keep you out.
- Genuine E-E-A-T: identified authors, expertise, trustworthy sourcing.
- Content that's timely and interest-driven, not just evergreen reference.
- Compelling titles that aren't clickbait - Discover updates have repeatedly punished sensationalism.
- Local relevance to the reader's country - increasingly decisive after the February 2026 Discover core update.
- Clean technical eligibility: indexable, fast, mobile-first.
Why the traffic swings
Because Google reallocates the feed constantly, Discover traffic can multiply overnight and vanish just as fast - especially around Discover-specific core updates. The defense isn't a trick; it's consistency: reliably publishing high-quality, well-imaged, locally relevant content so you're eligible often, instead of occasionally lucky.
If Discover matters to your business, I work it as a dedicated surface in the Discover & News service.
Key data
| Monthly users (2019) | 800M+ |
|---|---|
| Structured data required | No |
| GSC report history (months) | 16 |
The user figure is the last official public one, from 2019.
Sources: Google Search Central
Common mistakes and how to do it right
Discover punishes shortcuts more visibly than search. Learn from others' faceplants:
| Avoid | Do |
|---|---|
| ✗ Publishing with small or missing lead images | ✓ Large images over 1200px with max-image-preview:large |
| ✗ Clickbait titles to juice the feed | ✓ Compelling but honest titles - sensationalism gets demoted |
| ✗ Writing US-centric content for a local audience | ✓ Local relevance to the reader's country |
| ✗ Expecting steady Discover traffic | ✓ Building for consistent eligibility, not a single spike |
Tools I use and recommend
- Google Search ConsoleOrganic performance, coverage and field Core Web Vitals.
- PageSpeed InsightsCore Web Vitals and performance diagnostics (free, Google).
- Rich Results TestValidate structured data (free, Google).
- SemrushKeywords, site audit and competitor analysis.
Sources
Frequently asked questions
Can I target keywords for Google Discover?
No. Discover has no query - you become eligible through quality, images, freshness and local relevance, not keyword targeting. Discover & News service →
I lost Discover traffic overnight. Is that normal?
Unfortunately yes - Discover is volatile, especially around core updates. Recovery is about quality and local relevance. Feb 2026 update →