Google Discover Optimisation: The Mobile Feed Channel Most Sites Ignore

Google Discover is Google's personalised content feed, served in three primary surfaces: the Google app on iOS and Android, the Chrome mobile new-tab page, and the home-screen feed on most Android devices. Each user sees a feed personalised based on Google's understanding of their interests — derived from search history, app usage, location, content engagement, and signals from across the Google ecosystem. There is no query; users scroll a feed of Google-selected articles.

Discover is structurally different from search: surfacing is interest-based rather than query-based; the unit is the article card with a large image; visibility is per-user rather than per-query. The same article can be served to one user and not another based on personalisation. There is no rank tracking equivalent for Discover — Search Console reports impressions and clicks but doesn't expose per-user surfacing decisions.

For content sites, news publishers, and increasingly for brand-content sites in lifestyle, business, and technology categories, Discover often produces meaningful traffic with no explicit optimisation effort. In our managed engagements where we've measured carefully: Discover-attributed sessions (visible in Search Console's Discover report) typically produce 20–60% of total mobile Google traffic for sites that surface there at all. Sites that don't surface in Discover at all see effectively zero traffic from it.

The asymmetry matters. Sites surfacing in Discover gain a meaningful traffic channel that doesn't depend on ranking for specific queries, doesn't compete with AI Overviews, and doesn't suffer the CTR compression of traditional search. Sites not surfacing miss the channel entirely. Crossing the threshold from 'never surfaces' to 'surfaces regularly' is the single most valuable optimisation move and is achievable with deliberate effort over 60–120 days for most content sites.

Discover surfaces content the algorithm believes will interest the specific user being served. The patterns we see consistently associated with surfacing in our engagements:

• Strong, original imagery: Discover cards display a large image (1200x675 minimum for the high-resolution variant). Stock photography surfaces poorly; original photography and well-designed graphics surface materially better.
• Topical authority signals: sites with demonstrable depth on a topic (multiple recent articles, named expert authors, engagement signals) surface more than thin sites publishing one-off pieces in the topic.
• Recency weighting: Discover heavily favours recent content. Articles older than 14–30 days rarely surface unless they cover an evergreen topic the algorithm believes is currently relevant.
• Engagement signals from search and social: articles that performed well in search or got organic social engagement are more likely to be selected for Discover surfacing.
• Mobile-optimised reading experience: Discover users are on mobile by definition. Articles with poor mobile UX (intrusive interstitials, slow loading, layout shift) get suppressed even when other signals are strong.

The high-resolution Discover card variant (which Google preferentially serves on modern devices) requires a 1200x675 minimum image referenced via either schema.org/Article image property or by enabling the max-image-preview:large meta robots directive. Sites without this configuration default to the smaller card variant, which receives materially less engagement.

Implementation: ensure every article has a featured image at 1200x675 or larger, referenced in the Article schema's image property as a full URL (not relative), and either include max-image-preview:large in the robots meta or rely on Google's default behaviour for sites without robots restrictions. Without one of these signals Google won't enable the large card variant. We've seen this single fix double Discover impressions in audits where it was missing.

Discover heavily favours certain topic categories: news and current events, technology, business and finance, sports, lifestyle, food, travel, and entertainment all see strong Discover surfacing. Categories that perform poorly: most B2B services, narrow professional services (law, accounting, dental), local services without lifestyle angles, and anything that requires specialised professional knowledge to find interesting.

The mismatch is structural. Discover is a consumer entertainment surface; B2B and professional service content rarely fits that pattern. Sites in those categories often have poor Discover prospects regardless of optimisation effort. The honest evaluation: check Search Console's Discover report. If you have months of zero-impression data despite quality content, the topic itself is the issue and effort is better spent elsewhere.

Search Console's Discover report (Performance → Discover, available only to sites with sufficient Discover traffic to populate the report) shows total impressions, clicks, CTR, and per-page performance. The report appears for sites once they cross a minimum traffic threshold; sites with zero or minimal Discover surfacing won't see the report at all.

Track Discover impressions month-over-month as the primary indicator. Surfacing growth correlates with topical authority growth, image quality investment, and content velocity in the topic categories Discover favours. CTR varies wildly (3–15% is typical) and depends heavily on image quality and headline choice. Use the per-page report to identify which content patterns surface and double down on what's working in your specific account.