Google quietly gave 54 publishers control over their Discover profiles. Here’s what they did with it.

Google Discover has publisher profile pages. They live at profile.google.com/cp/ and appear when someone taps a publisher’s name on a Discover card. These pages aren’t new. They launched in August 2025 with the Follow button rollout, and by November 2025 Google’s documentation referred to them as “source overviews.”

For most of the 47,000+ publishers we monitored, the pages are auto-generated: a name, follower count, social links pulled from the Knowledge Graph, recent posts, and a footer label that reads “Profile generated by Google.”

Since March 2026, though, something changed for a small subset of publishers. A group gained access to enhanced profiles: custom banner images, a configurable links shelf, and the ability to pin posts (labeled “Pinned” in the publisher interface, formerly “Featured Posts”).

They also gained control over the order of their social links, website, and content tabs — something standard profiles don’t allow. On standard profiles, social links are sorted algorithmically by follower count, with the website listed last. On claimed profiles, the publisher decides.

The “Profile generated by Google” label also disappeared entirely, replaced by nothing — a quiet signal that the profile had been claimed.

There’s no public documentation explaining how to get access. No Search Console toggle. No application form. Google appears to have hand-selected participants for what is effectively an invitation-only pilot program.

We identified 54 publishers in this cohort. All are U.S.-based. All publish in English. And what they have — and haven’t — done with the feature over two months of monitoring reveals patterns every publisher should watch before the program scales.

How we found the 54

Our Profile Features Monitor tracks 46,926 publishers across seven languages: English, French, German, Italian, Spanish, Dutch, and Portuguese. To isolate the enhanced cohort, we filtered for publishers that showed persistent enhanced-profile signals across multiple snapshots: active links, full banner headers, or both.

The result: 54 domains with stable access to the enhanced profile surface. The composition of that group offers clues about Google’s intentions:

Tier	Publishers	Examples
National	15	WSJ, Fox News, NY Post, Newsweek, Inquirer
Regional Paper	13	Boston Globe, SFGate, CT Insider, Times Union
Local TV	14	KTLA, PIX11, MyFox8, WSMV, Atlanta News First
Lifestyle Brand	6	Delish, The Dodo, Country Living, House Beautiful
Specialty	6	Pew Research, The Athletic, Gothamist, Civil Beat

The skew toward local news and community publishers is striking and aligns with Google’s public emphasis on supporting local journalism. Nearly half the cohort — 27 of 54 publishers — consists of regional newspapers and local TV stations. National brands are included too, but they’re not the majority.

The two-tier profile system

Under the hood, Google operates two distinct profile architectures. Understanding the difference matters because this isn’t just a cosmetic upgrade. It’s a structural split.

Standard profile (99.9% of publishers):

• Auto-generated from public sources.
• “Profile generated by Google” label visible.
• No publisher control over content or layout.

Claimed profile (the 54 publishers):

• No generation label.
• Publisher can configure the banner, links shelf, and pinned post.
• Publisher controls the order of social links, website, and content tabs (standard profiles sort them by follower count).

This isn’t Search Console verification, structured data markup, or any existing publisher tool. It’s a separate, invitation-only system.

What the 54 publishers actually did

This is where it gets interesting. Access to a feature and its effective use are different. Here’s what the data shows across each configurable surface.

Banners: professional, deliberate, tier-predictive

Forty-one of the 54 publishers uploaded a banner image. The remaining 13 have the capability — a “prepared” state — but haven’t used it yet.

What stands out is the production quality. There are no amateur banners in the cohort. Every uploaded image reflects clear professional design investment.

Five distinct visual archetypes emerged:

• Brand-pattern: No photography, just the wordmark or abstract identity repeated as a tile. Pure prestige.
• Editorial content: The banner shows what the publisher covers. A food shot, a puppy, a stock chart.
• Local landmark: City skylines, local scenery, and regional identity anchors.
• Brand-statement: Curated collages with taglines or portfolio displays:
• Front-page archive: A grid of 12 iconic covers. Tabloid heritage as visual identity. Unique in the cohort.

Tier predicts archetype. National publishers cluster around brand-pattern banners. Local outlets lean into civic identity and city imagery. Lifestyle brands showcase their content directly.

One anomaly: The Athletic uploaded a solid black square — 656×656 pixels. Whether that reflects deliberate minimalism aligned with The Athletic’s dark UI or simply a broken upload is unclear. It’s the only non-image banner in the cohort.

The format split is revealing: 71% used square banners — likely Google’s recommended ratio — while 29% used wide landscape formats. None used portrait layouts. Based on CDN serving patterns, the minimum recommended resolution appears to be 512 pixels on the longest side.

Publishers that chose wide formats made deliberate design decisions: SecretNYC uses a manifesto-style collage, the New York Post uses a headline grid, and Barron’s uses a geometric pattern. Square appears to be the default safe option.

Links: local TV dominates, nationals ignore it

Thirty-three of the 54 publishers enabled the links feature. Of those, 31 added at least one link, for a total of 65 configured links across the cohort.

The content is overwhelmingly focused on on-site navigation: 85% of links point to the publisher’s own sections, weather pages, live streams, or app downloads. This functions more like a mini site navigation layer than a promotional surface.

The tier gap is enormous:

• Local TV: 31 links across 14 hosts (average 2.2 per publisher). Fox affiliates consistently shelve: Watch Live, Weather, Local News, Sub-region, Contact.
• National: 9 links across 15 hosts (average 0.6 per publisher). Most nationals didn’t bother.

Three outliers worth noting:

• PIX11 published “How to make PIX11 a preferred source on Google,” meta-promoting Discover follows from within the Discover profile itself.
• Gothamist funneled donations through `pledge.wnyc.org` with a purpose-specific utm_campaign=discover-profile tag.
• Fox Nation placed a direct subscription conversion link (“Subscribe to Fox Nation”) on what most publishers treat as a navigational surface.

Pinned posts (formerly Featured Posts): capability granted, rarely used

Fifty-two of the 54 publishers enabled the Pinned feature. Only 13 currently use it with an active pinned post.

Lifestyle brands were the strongest adopters: five of six had the feature active. Among national publishers, only 2 of 15 used it. The capability exists across nearly the entire cohort. Adoption does not.

About text: Wikipedia out, self-branding in

On standard profiles, the “About” section is auto-generated by Google, usually sourced from Wikipedia. On claimed profiles, publishers write their own.

Within the cohort, 38 of 54 use a custom-written description, while only 16 retain a Wikipedia-sourced version — a surprisingly low number for publishers of this size and prominence.

The tone splits cleanly by publisher tier.

• Local TV stations lean promotional (“Your trusted source for breaking news, accurate weather forecasts and local sports across Greensboro…” ).
• National and digital-native publishers stay more factual (“Gothamist is a website about New York City news, arts, events and food, brought to you by New York Public Radio”).
• One publisher takes a mission-driven approach: Delish — “you don’t have to know how to cook, you just have to love to eat!”

The implication for publishers preparing for this feature: once you claim the profile, you take control of the About section. It becomes your pitch on a Google-owned page.

Notably, the most visible publishers in the cohort chose factual descriptions over promotional copy.

UTM tracking: the blind spot

Only three of the 65 configured links include analytics parameters. Gothamist tagged its donation link with utm_campaign=discover-profile, making it the only publisher in the cohort treating the profile as a measurable acquisition channel.

The Philadelphia Inquirer instrumented two links, but one reused an Instagram bio campaign tag (mktg_acq_ig_organic_bio_offer), meaning Discover traffic from that link will be misattributed to Instagram in analytics.

The other 62 links have no tracking at all. In practice, 95% of the cohort has no way to measure whether profile links generate traffic.

Social platform priorities

On claimed profiles, publishers control the display order of social links and content tabs. Standard profiles don’t: Google sorts links algorithmically by follower count and places the website last. That means the ordering we observe on claimed profiles reflects deliberate editorial choices, not algorithmic defaults:

• Local TV stations list Facebook first: 86% (12 of 14). Zero list X/Twitter first.
• National publishers spread their bets: Facebook 33%, Instagram 20%, X 20%, YouTube 13%.
• Specialty/digital-native outlets lean Instagram-first (67%).

Concrete examples: Newsweek places YouTube first and Articles second. Delish leads with Website, followed by Instagram. These are active editorial decisions about which audience channel matters most.

The local TV finding is particularly notable. Despite news media’s historical reliance on X/Twitter, not a single local station in this cohort places it as their primary social link.

Sister-site coordination

For media groups with multiple properties in the cohort, setup patterns reveal whether profile management is centralized or handled locally:

• Hearst Connecticut, which has five papers in the cohort, shows near-identical configuration across all profiles. The links structure is the same, including a shared Hearst checkout funnel with publication-specific site IDs. The setup points to a centralized digital team managing profile operations across the group. Even so, each masthead still uses distinct banner art.
• Dow Jones, across The Wall Street Journal and jp.wsj.com, uses shared banner artwork: the same wordmark tile, confirmed through perceptual hashing. That points to brand coordination at the asset level.
• Everyone else Everyone else — including Fox affiliates, Dotdash Meredith properties, and the Fox News group — shows completely different setups across properties, even within owned-and-operated chains. Profile management appears to be handled locally rather than centrally.

The rollout is still active

Comparing snapshots #9 and #12 — taken 19 days apart — confirms this isn’t a frozen experiment. During that window, four publishers added banners (jp.wsj.com, New York Post, SecretNYC, and Everyday Health), one activated Links for the first time (New York Post), and jp.wsj.com (The Wall Street Journal’s Japanese edition) entered the cohort entirely.

No publishers lost features. The program is still expanding within the cohort, and new participants continue to appear.

The adoption paradox

We scored each publisher on a composite 0–6 scale, assigning one point for each of the following:

• Banner uploaded
• Links feature active
• Featured Posts active
• At least one configured link
• Four or more social platforms listed
• Any UTM tracking present

Nobody scored 6. The distribution:

Score	Publishers	%
2	22	41%
3	10	19%
4	14	26%
5	8	15%
6	0	0%

National publishers with the largest audiences are the least engaged with the configurable surface, with a mean score of 2.93. Most uploaded a banner and stopped there.

Local TV stations — despite having the smallest Discover footprints — are the most engaged, with a mean score of 3.57. Lifestyle brands score highest overall at 3.83, yet their Discover visibility trajectory is the flattest in the cohort.

And here’s the critical finding: feature adoption shows no correlation with visibility trajectory.

Across the cohort, the 180-day late/early capture ratio ranges from 0.23x for Prevention — down 77% — to 4.27x for NewsNation — up 327%. Variance is massive within every tier.

KTLA scores high on adoption, with seven links, a full banner, and active profile engagement, and grew 3.69x. But Delish also scores high and declined to 0.90x. MyFox8 configured five links and fell to 0.52x.

Publishers that fully utilized the configurable surface show no better visibility trajectory than those who used it minimally.

This feature gives publishers a controlled surface for branding and navigation, not a ranking lever. It’s a profile page, not an algorithm input.

What this means for publishers

The program is U.S.-only and invitation-only for now. Across the six other language markets we monitor — French, German, Italian, Spanish, Dutch, and Portuguese — we found zero enhanced profile deployments: not a single banner or configured link outside the English-language cohort.

But the underlying infrastructure is already in place. All 47,000+ publishers we track already have profile pages with follower counts, social links, and content feeds. The enhanced features sit on top of that existing architecture. Google isn’t rebuilding the system. It’s selectively unlocking capabilities within it.

If — or when — Google scales this, here’s how publishers should prepare:

• Audit your structured data now. Profile social links are pulled from your sameAs/JSON-LD markup. Errors there will carry over to your profile. Verify what Google will display before you’re given control.
• Design a banner. Use a square format (1:1 ratio) with a minimum resolution of 512px, and treat it as a professional brand asset. The 54 publishers in this cohort set a clear standard: there were no amateur images. Think about which archetype fits your brand: a wordmark tile for prestige brands, local landmarks for regional publishers, or content-driven imagery for vertical and lifestyle outlets.
• Plan your link strategy. The data suggests that section navigation and utility content — weather, live streams, and similar recurring destinations — drive the most engagement. Local TV stations treating the profile as a mini site navigation layer are the clearest power users. Decide now which five to seven links represent your most valuable entry points.
• Instrument from day one. Almost nobody in the current cohort tracks profile link performance. Adding a dedicated UTM campaign parameter — utm_campaign=discover-profile, for example — would put you ahead of 95% of the pilot group on attribution alone.
• If you’re a media group, decide your operating model. Should profile management be centralized or handled newsroom by newsroom? The cohort shows both models. Hearst Connecticut runs one coordinated setup across five papers, while Fox affiliates manage profiles independently at the station level. The important part is that the choice is deliberate — not something decided accidentally when individual newsrooms start receiving invitations.

Methodology

Data comes from the 1492.vision Profile Features Monitor, which tracks roughly 47,000 publishers across seven languages through recurring snapshots of profile metadata. The 54-publisher cohort was identified through persistent enhanced-feature signals observed across multiple snapshots between March and May 2026.

Visibility trajectories are based on proprietary capture data. All findings are descriptive only: the cohort reflects Google’s selection criteria, not a random sample, and this dataset does not support causal claims about feature impact.

The full analysis — including the complete 10-phase timeline, banner image gallery, snapshot-by-snapshot evolution, and tier-by-tier breakdowns — is available at 1492.vision/research/discover-publisher-profiles-en.