Surfaces third-party search applications inline on the SERP. The structural primitive for federated search and OneBox integration — third-party data sources appear as native SERP features.
Patent Overview
- Inventor
- Yossi Matias, others
- Assignee
- Google LLC
- Filed
- 2013
- Granted
- 2019-05-14
The Challenge
The Challenge
Some queries are best served by specialized third-party data. Flight prices, sports scores, stock quotes, weather, real-estate listings — all live in third-party systems. Surfacing them inline on the SERP requires federated query handling without breaking SERP latency or quality.
- Specialized Data Lives Off-Index — Many query types need specialized data Google doesn't directly index. Federation is required.
- Inline Integration Beats External Links — Sending users to third-party sites for simple info is friction. Inline SERP integration is better UX.
- Federation Must Be Fast — Third-party calls must fit within SERP latency budgets. Caching and timeout logic required.
- Source Quality Must Be Validated — Per third-party source, quality and reliability must be vetted. Unreliable sources damage SERP trust.
- Layout Integration Matters — Third-party blocks must integrate into SERP layout coherently. Inconsistent presentation breaks SERP experience.
Innovation
How The System Works
The system identifies queries that benefit from third-party data, runs federated queries to vetted third-party search applications, validates and integrates returned data into SERP blocks, and caches results for latency optimization.
- Identify Federation-Eligible Queries — Per query, identify whether third-party data would enhance the SERP.
- Select Applicable Third-Party Apps — Per query, select applicable vetted third-party search applications.
- Run Federated Query — Per app, run federated query in parallel with organic ranking.
- Validate Returned Data — Per response, validate data quality and freshness.
- Integrate Into SERP Block — Per response, integrate data into SERP block with consistent presentation.
- Cache For Latency — Per (query, app), cache responses for repeat-query latency optimization.
- Capture Engagement — Per block, engagement signals feed back into app selection and ranking.
Federation Brings Specialized Data Inline
The patent's load-bearing idea is that specialized third-party data belongs inline on the SERP when it improves UX. Federation, validation, integration, caching combine to make this work at SERP latency budgets.
Vetted Federation Plus Caching
Per third-party source, vetting ensures quality. Per federated query, caching ensures latency. The combination makes federation viable at scale.
- Vetted Third-Party Apps — Per app, quality vetted before federation eligibility.
- Per-Query Selection — Per query, applicable apps selected based on query intent.
- Latency-Optimized Caching — Per (query, app), caching reduces federation latency for repeat queries.
Technical Foundation
Technical Foundation
The patent specifies the federation-eligibility identifier, app selector, federated runner, response validator, SERP integrator, cache, and engagement-feedback loop.
- Federation-Eligibility Identifier — Per query, identifies federation eligibility.
- App Selector — Per query, selects applicable third-party apps.
- Federated Runner — Runs federated queries in parallel.
- Response Validator — Validates returned data quality and freshness.
- SERP Integrator — Integrates responses into SERP blocks.
- Cache — Caches responses for latency optimization.
The Process
The Process
Per query, federation pipeline runs alongside organic ranking.
- Receive Query — Query arrives.
- Identify Federation Eligibility — Eligibility classifier runs.
- Select Apps — Applicable apps selected.
- Run Federated Queries — Federated queries run in parallel.
- Validate Responses — Response data validated.
- Integrate Into SERP — Responses integrate into SERP blocks.
- Cache And Track Engagement — Responses cached; engagement tracked.
Quality Control
Quality Control
Federation depends on third-party quality. The patent specifies safeguards.
- App Vetting — Third-party apps quality-vetted before federation eligibility.
- Response Validation — Per response, data quality and freshness validated.
- Latency Budget Enforcement — Per federated query, latency budget enforced. Timeouts respected.
- Cache Freshness — Per (query, app), cache freshness validated. Stale data invalidated.
- Engagement Monitoring — Per app, engagement monitored. Low-engagement apps deprioritized.
Real-World Application
Third-party federation powers many specialized SERP features: flight prices, weather, stocks, sports. The pattern of vetted-federation plus integration plus caching is the architectural template.
- Vetted apps Federation Sources — Third-party apps vetted for quality.
- Per-query Selection Granularity — Per query, applicable apps selected.
- Cached Latency Strategy — Responses cached for repeat-query latency optimization.
Why Structured Data Sources Win Federation Eligibility
Federation favors structured, machine-readable third-party sources. Sites exposing structured data via APIs and Schema.org markup are federation-ready in ways unstructured sites are not.
Why Reliability Compounds For Federation Sources
Federation eligibility is a long-term relationship. Sources delivering reliable, fresh, well-validated data become preferred federation partners over time.
<\/section>What This Means for SEO
What This Means for SEO
This patent surfaces specialized third-party data inline on the SERP through vetted federation, validation, and caching. SEO implication: structured, machine-readable, reliably-fresh data sources are federation-ready and become preferred partners over time.
- Structured Data Wins Federation Eligibility — Federation favors structured, machine-readable third-party sources. Sites exposing data via APIs and Schema.org markup are federation-ready in ways unstructured sites are not, which is the entry requirement for inline surfacing.
- Reliability Compounds Into Partnership — Federation eligibility is a long-term relationship. Sources delivering reliable, fresh, well-validated data become preferred federation partners over time, so consistency builds standing.
- Source Quality Is Vetted Before Eligibility — Third-party apps are quality-vetted before they can be federated. Being a credible, accurate data source is a prerequisite, so federation rewards genuine reliability rather than mere availability.
- Freshness Is Validated Continuously — Returned data is validated for quality and freshness, and stale cache is invalidated. Keeping your data current is what keeps you eligible, since outdated responses fail validation.
- Per-Query Selection Rewards Relevance — Applicable apps are selected per query based on intent. Offering data that clearly matches specific query intents, like prices, scores, or availability, is what gets you selected for those queries.
- Engagement Deprioritizes Weak Sources — Per-app engagement is monitored and low-engagement apps are deprioritized. Federated data that genuinely serves users holds its placement; data users ignore loses it.
- Latency Discipline Is Required — Federated queries must fit within SERP latency budgets, with timeouts respected. Fast, dependable response times are part of being a viable source, so performance is a federation requirement, not a nicety.