Claim
The March 2026 Google core update completed a structural shift that several earlier updates pointed toward: comparison sites, OTAs, job boards, and review aggregators face shrinking visibility regardless of content quality, while authoritative brand-owned domains gain. Even YouTube — historically a winner — saw the single greatest visibility loss of the update. The pattern is not a quality filter that aggregators happened to fail; it is a structural reweighting that says: middleman content layers between the brand and the user no longer carry the same weight. Owned-brand authority — the brand's own domain, original research, named experts, first-party voice — is the asset that survives.
Mechanism
Search has been moving in this direction for years; the March 2026 update made the move legible. The mechanism is a combination of two effects. (1) AI surfaces (AI Overviews, ChatGPT/Claude/Perplexity) intercept more of the top-of-funnel queries that aggregators historically captured, so the SERP itself reweights toward outputs that complement rather than compete with AI synthesis — and AI synthesis already loves brand-owned, primary sources. (2) Google's own quality signal evolution rewards demonstrated brand authority (E-E-A-T, named authorship, original research, first-party data) and penalises content that exists primarily to capture clicks on someone else's category. Aggregator economics depend on being the layer between the user and the brand; when the user can talk to the AI directly and the SERP rewards the brand directly, the aggregator's role evaporates.
Conditions
Holds when:
- The category has aggregator presence to lose share from (B2B SaaS, ecommerce, travel, finance, real estate).
- The brand has any authentic E-E-A-T signal to point to — named people, real research, first-party data, customer outcomes.
- The team has authority to invest in long-cycle owned-brand work (research, named-author content, video) rather than short-cycle SEO.
Fails when:
- The brand is genuinely commoditised and has no E-E-A-T to develop (some commodity infrastructure).
- The aggregator's value is genuinely useful (price comparison in a real-time category) and the brand can't replicate it.
- The team interprets "brand authority" as "more branded content" without changing what the content actually is — the structural shift is rewarding signals, not output volume.
Evidence
The March 2026 core update analysis: comparison sites, OTAs, job boards, review aggregators saw visibility losses that didn't correlate cleanly with content quality. YouTube — long a structural winner — lost more SISTRIX visibility points in this update than any other domain category. Brand-owned domains in finance, healthcare, and SaaS gained.
— Lily Ray, Google March 2026 core update — winners and losers analysis, 2026-04-30.
Two operators independently arrived at the same conclusion the same week: Aleyda Solis's 87.6M-visit global AI search study (Apr 29) found per-vertical concentration patterns favouring brand-owned domains; Kevin Indig's Growth Memo (May 4) framed it as judgement and brand authority being the non-compressible parts of the workflow.
Signals
- Investment moves from aggregator placements (sponsored review sites, listing platforms) to owned-brand assets (research, named authors, video).
- E-E-A-T signal audits — named author bylines, credentials, original data, source citations — become regular review items.
- The brand owns the high-intent comparison content rather than competing with aggregators on it.
- Internal dashboards distinguish brand-domain visibility from aggregator-mediated visibility.
Counter-evidence
- Aggregators won't disappear; in some categories (price-comparison, real-time inventory) they remain useful and continue to capture meaningful traffic.
- The shift is happening at different rates per vertical; some categories haven't fully transitioned yet, and over-rotating to owned-brand investment in those can be premature.
- Brand authority isn't infinite — operators with no real authority to develop will not build it through tactical moves alone, and chasing the signal without the substance is its own failure mode.
Cross-references
- A single seeded fake claim can self-confirm in AI Overviews — Lily Ray's parallel finding on the AI surface side: brand-protection now requires monitoring for hallucinated brand claims.
- Citation rate and mention rate are different metrics; comparative content closes the gap — Indig's data on which content types AI surfaces actually name brands in.
- AEO is a GTM capability, not an SEO experiment — Voje's framing of why this work belongs to PMM: brand authority traces to positioning, not keywords.
- Building costs collapsed; judgement didn't — the squeeze is on positioning, not production — Indig's parallel framing: same week, same call from a different angle.