Per-user personalized navigation through search engine. Personalization-driven result re-ranking — each user's history shapes which results best serve them.
Patent Overview
- Inventor
- Susan T. Dumais, others
- Assignee
- Microsoft Corporation
- Filed
- 2010
- Granted
- 2014-08-05
The Challenge
The Challenge
Per user, search history reveals preferences. Per query, personalized navigation surfaces results aligned with the user's prior engagement patterns. Personalization must work without violating privacy.
- Users Have Distinct Preferences — Per user, search and navigation patterns differ.
- Per-User History Reveals Intent — Per user, history shapes what's relevant for them.
- Personalization Improves Relevance — Per query, personalized results match individual user need.
- Privacy Must Be Preserved — Per user, personalization respects privacy.
- Personalization Must Be Bounded — Per query, personalization shouldn't filter out objectively-relevant results.
Innovation
How The System Works
The system captures per-user search and navigation history with consent, infers per-user preferences, modulates ranking by per-user signals, and respects privacy throughout.
- Capture User History With Consent — Per user with opt-in, history captured.
- Infer User Preferences — Per user, preferences inferred from history.
- Receive Query — Query arrives.
- Apply User Preferences — Per (user, query), preferences modulate ranking.
- Rank Personalized Results — Per user, ranking applied.
- Bound Personalization — Per query, personalization bounded to prevent filter bubbles.
- Privacy Preserve — Per user, signals handled with privacy.
User History Drives Personalization
The patent's load-bearing idea is that per-user history shapes per-user ranking. Personalized navigation surfaces results aligned with individual preferences.
Per-User Preference Modeling
Per user, preferences inferred from history. Per query, preferences modulate ranking.
- Consent-Based History Capture — Per user, opt-in capture.
- Preference Inference — Per user, preferences inferred.
- Bounded Personalization — Per query, personalization bounded.
Technical Foundation
Technical Foundation
The patent specifies the history capturer, preference inferrer, ranking modulator, bound applier, and privacy layer.
- History Capturer — Per user with consent, history captured.
- Preference Inferrer — Per user, preferences inferred.
- Ranking Modulator — Per (user, query), preferences modulate ranking.
- Bound Applier — Per query, personalization bounded.
- Privacy Layer — Privacy safeguards on signals.
- Recalibration — Preferences refresh as behavior evolves.
The Process
The Process
Per query, personalized ranking runs in real time.
- User Opts In — Consent captured.
- Capture History — Per user, history accumulates.
- Infer Preferences — Per user, preferences inferred.
- Receive Query — Query arrives.
- Modulate Ranking — Per (user, query), ranking modulated.
- Apply Bounds — Personalization bounded.
- Return Results — Personalized results returned.
Quality Control
Quality Control
Personalization must avoid filter bubbles. The patent specifies safeguards.
- Privacy Preservation — Per user, signals handled with privacy.
- Personalization Bounds — Per query, modulation bounded.
- Filter-Bubble Prevention — Per query, diversity maintained.
- User Control — User can review, edit, opt out.
- Continuous Recalibration — Preferences refresh.
Real-World Application
Personalized navigation is foundational across modern search. The pattern of consent-based history capture plus bounded personalization underpins how engines balance personalization against diversity.
- Per-user Personalization Granularity — Each user has personalized ranking.
- History-driven Inference Source — Per user, history shapes preferences.
- Privacy-preserved Architecture — Privacy safeguards on signals.
Why Returning Visitors Build Preference Signal
Per user, history reveals which sites earn return engagement. Sites earning return visits accumulate per-user preference signal compounding across personalized ranking.
Why Multi-Surface Audience Matters
Per user, personalization shapes which sites surface. Building genuine audience preference compounds across personalized ranking in ways pure-SEO optimization cannot.
<\/section>What This Means for SEO
What This Means for SEO
Per-user search and navigation history is captured with consent and used to bound-personalize ranking toward sites a user has engaged with. SEO implication: earning return visits builds a per-user preference signal that pure on-page optimization cannot replicate.
- Return Visits Build Preference Signal — History reveals which sites earn repeat engagement, and those sites surface more for that user. Becoming a destination users come back to compounds across personalized ranking. Cultivate loyalty, not just one-time clicks.
- Audience Beats Optimization Alone — Personalization rewards genuine user preference that optimization cannot fake. Building a real audience that chooses you produces a signal pure keyword work cannot, so invest in brand and reader relationship.
- First Impressions Seed The Loop — Per-user preference starts from prior engagement, so the initial satisfying visit is what earns the return that then boosts you. Make the first visit good enough to start the personalization flywheel.
- Personalization Is Bounded, Not Absolute — The system deliberately avoids filtering out objectively relevant results. Personalization tilts the ranking, it does not guarantee a slot, so you still need baseline relevance and quality to be in contention.
- Brand Recognition Helps You Get Chosen — When a user has engaged with you before, you are favored among comparable results. Strong, memorable brand presence increases the odds users re-engage and that preference accrues to you.
- Consent-Gated Capture Means Genuine Engagement Counts — History is captured with consent and reflects real navigation. There is no shortcut signal to manufacture; the lever is producing experiences worth returning to.
- Multi-Surface Presence Reinforces Preference — Personalization shapes which sites surface across a user's activity. A consistent presence across the surfaces your audience uses reinforces the preference signal and keeps you in their personalized rotation.