Google’s Preferred Sources rolled out worldwide on April 30, 2026, and users are twice as likely to click through to a publisher they have marked. More than 200,000 sites have already been selected. The problem: GA4 buckets every one of those clicks as google / organic, and Search Console does not separate Preferred-Source impressions from algorithmic ones. The lift is real. Your dashboards cannot see it.

What Preferred Sources Actually Is

Preferred Sources is a Google Search feature that lets users explicitly mark publishers they want surfaced more often in Top Stories, Discover, and personalized news experiences. The feature was first introduced in 2025 in a limited rollout across selected regions and languages. On April 30, 2026, Google made it globally available across every supported language — bringing EU, LATAM, APAC, and MENA markets into the loyalty layer for the first time.

The two stats Google attached to the announcement are the ones that matter for any WooCommerce content publisher: users who mark a source as preferred click through to it twice as often, and more than 200,000 unique sites had already been marked as preferred by users globally at the worldwide rollout.

Translation: Google has now wired user-declared loyalty into its ranking layer for Discover and Top Stories. If you have been investing in AEO content, brand recall, or any kind of repeat-reader strategy on WooCommerce, this is the surface where that investment is supposed to pay off.

The Measurement Gap You Inherit on Day One

Here is what your reporting stack actually does with that lift.

GA4 aggregates everything into google / organic. The google / organic source/medium combination is the default GA4 bucket for traffic from Google Search. It contains standard SERP listings. It contains AI Overview citations. It contains Top Stories clicks. It contains Discover clicks. And as of April 30, it contains Preferred Source clicks. GA4 has no built-in dimension to separate them. There is no parameter, no custom dimension, no report inside the product that distinguishes a user who clicked because Google guessed they would from a user who clicked because they explicitly told Google to surface this site more.

Search Console aggregates Preferred-Source impressions into the baseline. The queries report shows total clicks and impressions for a query. The Discover performance report shows total Discover clicks. Neither report segments out the impressions that were elevated because a user marked you as a Preferred Source. From inside Search Console, the lift simply raises every number a little, with no way to attribute the cause.

Every Preferred Sources article tells publishers how to encourage users to mark them. None of them tell you what happens after. When the lift arrives, you have no native instrumentation to see it. You can be doing everything right — winning the loyalty layer, doubling click-through on your most engaged readers — and your dashboard will report the gain as a small, unattributable rise in google / organic.

You may be interested in: Google AI Mode Hit 75 Million Daily Users and Then Buried Itself Inside the Organic Search Bucket

Why This Lift Is Worth More in 2026 Than It Would Have Been in 2024

The organic baseline has already collapsed. AI Overviews are eating roughly 58% of click-throughs on position-one listings for WooCommerce content, which means the baseline that Preferred-Source elevation sits on top of is dramatically smaller than it was eighteen months ago. The math cuts two ways. Every Preferred-Source click is more valuable per visitor because the algorithmic baseline has thinned. And every Preferred-Source click is harder to detect, because it is being added to a noisier, smaller, faster-changing organic bucket.

This is the broader pattern. Google has been launching new surfaces — AI Mode, AI Overviews, Discover, Top Stories, Shopping Graph, now Preferred Sources — and consistently routing the resulting traffic into the same undifferentiated google / organic dimension in GA4. The surfaces multiply. The attribution stays flat. For a WooCommerce store that has built brand-loyalty content as its growth strategy, this is the structural problem. The traffic is real. The measurement framework is not.

The Proxy Framework: What You Can Actually Measure

Since GA4 and Search Console will not segment Preferred-Source traffic for you, the only working approach is an indirect one — capture the behavioural signature that loyalty produces, on infrastructure you control.

Two signals matter.

Returning-visitor signal. A user who marks you as a Preferred Source is, by definition, declaring an intent to return. The act of marking is itself a loyalty event. Users who do it are statistically more likely to come back, visit deeper pages, and arrive directly rather than through search. First-party session data on returning visitors — captured on your own subdomain before browser restrictions and aggregation rules apply — is the closest behavioural proxy you have for the size of your preferred-source audience.

Brand-search uplift. Users who add a publisher as preferred are also more likely to search the brand name directly. Branded-search volume — visible in Search Console as queries containing your brand — moves in step with loyalty growth in a way that pure algorithmic traffic does not. Watching branded queries as a separate series, against total google / organic, isolates the part of your organic growth that is being driven by user-declared intent rather than algorithm whim.

Neither signal is a precise measurement of Preferred-Source clicks. Both are directional indicators of the underlying loyalty layer that Preferred Sources reflects. Together, watched over weeks, they give you something Search Console and GA4 will not: a defensible read on whether your AEO and brand-recall investment is producing the loyalty Google is now rewarding.

You may be interested in: AI Overviews Are Eating WooCommerce Click-Throughs by 58% on Position-One Listings

Why First-Party Server-Side Is the Only Place This Proxy Works

The proxy framework above relies on data that client-side analytics increasingly cannot capture cleanly. Returning-visitor identity is what Safari’s Intelligent Tracking Prevention specifically shortens to seven days of first-party cookies. Direct traffic is what ad blockers misclassify when scripts fail to load. Branded search arrives at your site through redirects, app contexts, and AI surfaces that strip referrers. The signal exists. Client-side instrumentation cannot reliably hold onto it.

Server-side, first-party tracking captures session continuity before the browser layer interferes. Events flow through your own subdomain — not a third-party tag — so ad blockers do not see them as foreign requests. First-party cookies set by your own server are not subject to the ITP seven-day limit. Returning visitors stay identified across sessions. Branded-search referrers are preserved.

Here Is How You Actually Do This

Build the proxy on first-party data captured server-side from your own subdomain. Transmute Engine™ is a dedicated Node.js server that runs on your own subdomain (e.g., data.yourstore.com), receives events from WordPress via the inPIPE plugin, and routes them simultaneously to GA4, BigQuery, and your other destinations — with full session continuity for returning visitors and clean branded-search referral data preserved at the server before the browser can erode it. You can run the returning-visitor and brand-search proxy off the BigQuery stream directly, week over week, without depending on GA4’s aggregated source/medium bucket to do the work it has just confirmed it will not do.

Key Takeaways

• Preferred Sources doubles click-through for marked publishers — and rolled out worldwide on April 30, 2026 across every supported language.
• 200,000+ sites had already been marked as preferred by users globally at the moment of worldwide rollout.
• GA4 aggregates Preferred-Source clicks into google / organic with no native way to segment.
• Search Console does not separate Preferred-Source impressions from baseline query or Discover performance.
• The working proxy is returning-visitor signal plus brand-search uplift, captured first-party on your own server before browser restrictions and aggregation rules collapse the signal.

Frequently Asked Questions

The lift is real. The default dashboards will not show it. Build the proxy first-party, on your own subdomain. seresa.io