The foundational Knowledge Panel display patent. Defines how the entity card appears alongside SERP results, what facts it surfaces, and how it interacts with the result list to form a unified entity-aware search experience.
Patent Overview
- Inventor
- Jeromy William Henry, others
- Assignee
- Google LLC
- Filed
- 2013-12-19
- Granted
- 2016-09-27
- Application Number
- US 14/135,221
The Challenge
The Challenge
Entity-rich queries return ten blue links that all describe the same entity slightly differently. Users have to scan the result list to assemble facts. The system needed a way to surface entity facts directly while still showing relevant documents, without crowding either out.
- Document Lists Force Manual Fact Assembly — When a user queries a person, place, or company, the result list returns articles about the entity. The user still has to read articles to assemble basic facts. A direct panel could deliver them at a glance.
- Panel Must Not Crowd Result List — Surfacing the panel cannot replace the documents users still want. The layout must accommodate both panel and results without forcing one out.
- Entity Facts Must Be Authoritative — Wrong facts in the panel cost user trust quickly. Source authority and fact provenance are critical.
- Panel Composition Varies By Entity Type — A person panel needs birth date, occupation, photo; a restaurant panel needs hours, phone, address; a company panel needs founders, employees, products. The composition must adapt to entity type.
- Interaction Must Stay Lightweight — Users should be able to scan the panel without leaving the SERP. Deep exploration is a click away, not the default flow.
Innovation
How The System Works
When the query resolves to an entity with sufficient confidence, the system retrieves the entity record from the knowledge graph, selects facts appropriate to the entity type, composes them into a panel layout, and renders the panel alongside the standard result list.
- Resolve Query To Entity — Query annotation identifies entity candidates. Above-confidence-threshold entities trigger panel display.
- Fetch Entity Record — From the knowledge graph, fetch the full structured record for the resolved entity: name, type, description, attributes, related entities, images.
- Determine Entity Type — Entity type (Person, Place, Organization, Product, Event) drives which panel template applies and which facts are surfaced first.
- Select Facts For Display — Per template, pick the facts that fit the panel slots. Time-sensitive facts (hours, prices) prioritize for immediate-intent queries.
- Compose Panel Layout — Fill the template with selected facts. Image at top, name and description, key attributes, related entities, source citations.
- Render Alongside Results — The composed panel renders in the right-rail or top-of-SERP slot, alongside the document result list. Both surfaces coexist.
- Capture Feedback — Panel clicks, dwell, expansions feed back into facet selection and ranking. The system learns which facts users care about for which entity types.
Entity Panel As First-Class SERP Element
The patent's load-bearing idea is to elevate entity facts from buried prose in documents to a first-class panel element rendered alongside results. The SERP becomes entity-aware, not just document-aware.
Facts At A Glance Plus Documents To Explore
The panel delivers facts immediately; the document list supports exploration. Both surfaces serve different user needs and coexist on the SERP.
- Entity Resolution Gates Display — Panels appear only when the query confidently resolves to a known entity. Ambiguous or non-entity queries skip the panel.
- Type-Specific Templates — Person, Place, Organization, Product. Each entity type has its own template with type-appropriate fact slots.
- Alongside, Not Replace — The panel coexists with the result list. Users get both surfaces; the layout accommodates each.
Technical Foundation
Technical Foundation
The patent specifies the entity resolution gate, the entity record store, the type-specific templates, the fact selector, the panel compositor, and the SERP layout integration.
- Entity Resolution Gate — Query annotator outputs candidate entities with confidence. Above-threshold confidence triggers panel pipeline; below-threshold suppresses.
- Entity Record Store — Knowledge graph stores per-entity structured records. Fast lookup is essential since the panel must compose within the SERP latency budget.
- Type-Specific Templates — Each entity type has its own template defining slot layout, fact priorities, and rendering. Templates are versioned and updated centrally.
- Fact Selector — Per entity record and template, picks the facts that fit panel slots. Time-sensitive facts get priority for immediate-intent queries.
- Panel Compositor — Fills the template with selected facts. Handles missing-fact fallbacks, image sizing, and citation rendering.
- SERP Layout Integration — Panel renders in the right-rail slot on desktop, top-of-results on mobile. Layout adapts to viewport and surface.
The Process
The Process
Panel composition runs in the query path. Entity resolution, record fetch, template fill, and rendering all complete within the SERP latency budget.
- Receive Query — Query enters the dispatcher. Entity annotation runs first.
- Resolve Entity — Annotator identifies candidate entities. High-confidence cases proceed to panel composition; low-confidence cases suppress the panel.
- Fetch Entity Record — Knowledge graph store returns the full structured record. Caching keeps lookup fast.
- Apply Template — Entity type determines template. Template defines slot layout and fact priorities.
- Select Facts — Fact selector picks per template. Facts not fitting the layout are dropped or relegated to expansions.
- Compose And Render — Compositor fills the template, handles missing fields, renders the panel HTML alongside standard SERP.
- Capture Interactions — Panel interactions log per entity per user. Feedback feeds template refinement.
Quality Control
Quality Control
Wrong panels cost user trust quickly. The patent specifies safeguards.
- Confidence Threshold Calibration — Panel display only when entity resolution confidence exceeds threshold. Wrong panels are worse than missing panels.
- Source Authority Weighting — Facts in the panel come from authoritative sources. Low-authority claims are excluded or labeled.
- Time-Sensitive Fact Validation — Hours, prices, schedules are validated for freshness. Stale facts are suppressed rather than displayed misleadingly.
- Disambiguation For Common Names — Names shared by multiple entities trigger disambiguation flow. Wrong-entity panels are worse than no panel.
- User Feedback Channel — Reports of wrong facts feed the upstream knowledge graph quality pipeline. Common errors get corrected at source.
Real-World Application
The Knowledge Panel is one of the most visible features of modern Google Search. This patent is the foundational architecture behind the panel as users experience it daily across queries for people, places, organizations, products, and events.
- Entity-resolved Display Trigger — Panels appear only when the query confidently resolves to an entity. Other queries see the standard result list.
- Type-specific Template Selection — Each entity type has its own template. The panel composition adapts to what the entity is.
- Alongside Layout Relationship — The panel coexists with the result list. Users get both immediate facts and documents to explore.
Why Schema Markup Compounds Panel Visibility
The panel reads from the knowledge graph. Pages with explicit Schema.org markup contribute clean entity data that the graph ingests. Pages with strong markup coverage compound the visibility of the entities they cover.
Why The Panel Is A Zero-Click Surface
Users who get their answer from the panel do not click into documents. This reshapes traffic distribution for entity queries: the panel captures attention; documents see less click-through. Brand investment in being the panel's authoritative source becomes a structural advantage.
<\/section>What This Means for SEO
What This Means for SEO
The patent defines the Knowledge Panel: when a query resolves to an entity, the system pulls knowledge-graph facts into a panel rendered beside the result list. SEO implication: entity facts become a first-class zero-click surface, so being the authoritative, well-marked source behind an entity is a structural advantage.
- Schema Markup Feeds The Panel — The panel reads from the knowledge graph, which ingests clean entity data including Schema.org markup. Pages with strong, accurate markup contribute the data that populates panels, compounding visibility for the entities they cover.
- The Panel Is A Zero-Click Surface — Users who get their answer from the panel do not click documents, reshaping traffic for entity queries. Plan for entity queries to yield fewer clicks, and make being the panel's authoritative source the goal rather than just ranking below it.
- Entity Confidence Gates The Panel — The panel appears only when the query resolves to an entity with sufficient confidence. Consistent, unambiguous naming and structured data across the web raise the system's confidence in your entity, making the panel more likely to trigger correctly.
- Fact Accuracy Is Brand Work — The panel selects facts appropriate to the entity type and presents them as authoritative. Ensuring the underlying knowledge-graph and source data about your entity is accurate and complete is essential brand hygiene, because errors surface prominently.
- Panel And Documents Coexist — The panel delivers facts while the document list supports exploration; both serve different needs on one SERP. Optimize for both: be the entity's panel source and rank the supporting documents users click to explore deeper.
- Be The Canonical Entity Source — The graph picks authoritative sources for each fact. Establishing your site or profiles as the canonical source for facts about an entity (via consistent, structured, widely-corroborated data) is how you influence what the panel shows.
- Entity Type Drives Fact Selection — Facts surfaced depend on entity type. Marking up the correct type (Person, Organization, Place, Product) ensures the panel pulls the relevant fact set, so type accuracy in your markup directly shapes the panel's content.