Where recommendations appear
Recommendation systems may organize a Following feed, discovery feed, search results, suggested people or businesses, groups and spaces, events, courses, marketplace listings, sponsorship opportunities, notifications, and related-content modules. Not every surface is personalized, and available feeds may change.
Principal ranking signals
Depending on the surface, principal signals may include whether you follow or interact with an account; the recency and frequency of activity; shared connections; topics, keywords, categories, format, language, and approximate location; prior views, clicks, saves, reactions, comments, hides, purchases, or reports; listing relevance, availability, service area, quality, and seller history; and technical, integrity, or safety indicators.
How signals are weighed
Systems generally prioritize relevance, timeliness, relationship strength, usefulness, diversity, quality, and safety. A signal’s weight varies by surface and context. Recent activity may matter more in a feed, text relevance in search, and availability or buyer criteria in marketplace results. No signal guarantees placement, distribution, engagement, or sales.
Negative and safety signals
Spam indicators, repeated hides, blocks, credible reports, misleading metadata, manipulated engagement, low-quality duplication, policy violations, unavailable products, unresolved transaction issues, or safety risks may reduce distribution or remove eligibility for recommendation. Enforcement decisions are governed by the Content Moderation, Reporting, and Appeals Policy.
Your controls
You can influence recommendations through available actions such as follow, unfollow, connect, save, hide, mute, block, report, or mark “not interested.” Profile, language, location, and topic preferences may further shape results. If a chronological or less-personalized feed is offered, you may select it using the controls shown in that feature.
Sponsored content and paid placement
Paid or sponsored content will be labeled where required. Eligibility, audience selection, bid or budget, predicted relevance, quality, and safety may affect delivery. Marketplace placement may also reflect paid promotion, which will be identified. Payment for placement does not override policy, safety, or eligibility requirements.
AI-generated and synthetic content
Content produced or materially altered by AI may be labeled, deprioritized, or made ineligible for certain recommendations when required for authenticity, safety, or user control. Members must preserve provenance signals and make disclosures required by the AI Use and Generated Content Policy.
Integrity and manipulation
Do not buy or sell fake engagement, coordinate deceptive interaction, create undisclosed bot networks, stuff listings with irrelevant terms, or otherwise manipulate recommendation or search systems. We may discount inauthentic signals and restrict associated accounts, content, or transactions.
Testing and changes
We test ranking approaches to improve relevance, safety, fairness, and performance. This may change the signals, weights, interfaces, and controls described here. We will update this Policy when principal practices materially change and provide additional disclosure where required.
Contact
Questions about feeds, search, or recommendation controls may be submitted through FocalShift Contact.