LinkedIn, in plain words

A section of your LinkedIn profile where you introduce yourself and tell your professional story.

A section on a member’s profile that shows verification information, such as checks of their identity, workplace or educational institution. It also shows when the profile was created and last updated.

A score LinkedIn’s automated systems give each signup attempt to assess the risk of a fake account. Low scores allow signup, medium scores trigger checks to confirm a real person, and high scores block account creation.

How a viewer’s reaction to a post can lead other people to see it and engage with it, as considered in a social network feed-ranking technique patented by Microsoft.

A process that chooses which ads to show on LinkedIn, using predictions of how likely a member is to click an ad.

A public, searchable collection of ads that have run on LinkedIn.

Controls that let you manage your advertising preferences and how data collected on LinkedIn and from other sources is used to choose the ads you see.

Low-effort content that is likely made with AI and may sound polished but lacks a clear point of view, an original perspective, or substance. It can feel generic, repetitive, or recycled.

Personalized outreach messages that recruiters can write with help from LinkedIn’s AI to contact potential candidates.

AI-generated starting material for collaborative articles, which are LinkedIn-published articles with members’ insights on professional topics. A LinkedIn team oversees the development of these conversation starters.

An artificial intelligence tool that helps you tailor your job applications directly on LinkedIn, reducing the need for third-party services.

Unequal treatment of groups by LinkedIn’s automated systems, and efforts to detect and reduce it. LinkedIn checks how Feed content is distributed and ranked for bias, including differences by gender, to support fairer outcomes for members.

Long-form pieces you publish on LinkedIn using its publishing tool. You can publish them as yourself or a LinkedIn Page, either on their own or as part of a newsletter.

Clues about how genuine or trustworthy a profile or Page is, such as verification status and how many members block it. LinkedIn may use these clues to set posting limits.

Automated tools that identify content that breaks LinkedIn’s rules and help flag or remove it from the platform.

A Microsoft patent describes a bidding strategy as a choice about what an ad campaign’s bids prioritize, such as the cost of clicks, ad displays or gaining customers.

A way of estimating members’ gender using two categories. LinkedIn combines these estimates with gender information members provide to check its algorithms for gender bias, including in how Feed content is distributed and ranked.

An advertising program that lets brands place video ads before videos from trusted publishers and creators in members’ LinkedIn feeds. It was previously called the Wire Program.

A technique described in Microsoft patents that controls how quickly an advertising campaign spends its budget and can limit when its ads are shown.

An AI technique LinkedIn's researchers have described for predicting how likely a member is to click an ad in their home feed, using their profile and past ad views and actions.

The first step in LinkedIn’s recommendation process, which narrows a large pool of possible posts or people to a smaller set. That set then gets more detailed scoring to decide what to show you.

A section of My Network that highlights reasons to start conversations with your existing connections, such as a new job or a work anniversary.

Replies people leave on posts or articles to join a conversation. You can control who comments on your posts or turn comments off.

An organization’s presence on LinkedIn, where it can share posts, articles, and other updates. Members can follow it to receive updates and discover the organization’s employees and job openings.

A number showing how likely a system thinks a fact or match is to be correct. Microsoft patents describe confidence scores for inferred profile facts and for matching search words to known skills or companies.

A cutoff for how certain a system must be before linking a household watching TV with an online user. A Microsoft patent describes this technique for extending TV ad campaigns online.

A technique LinkedIn's researchers have described for narrowing thousands of Feed posts from your network to a few hundred before final ranking. It considers how close you are to each post's author and what the post is about.

Members you know and trust who you’ve connected with on LinkedIn. You become first-degree connections, can message each other directly, and can see each other’s posts and updates in your feeds.

A LinkedIn Page administrator role that lets you publish updates and articles on the Page’s behalf.

The process LinkedIn uses to detect and review content that may break its policies, using automated systems and human reviewers, and remove content that violates those policies.

An assessment of how much value content adds to professional conversations on LinkedIn. It helps LinkedIn decide what to show in the feed, with generic or repetitive content less likely to be widely shown.

A process that scores and orders posts in your LinkedIn feed using information about you, your interests and your activity, aiming to show content you are likely to find useful or engaging.

A six-week program for selected LinkedIn creators that provides coaching, support, a financial award, and early access to tools to help them grow their audience and engage their community.

Tools that show how your LinkedIn content performs and who follows you, including how often posts appear on screen, how people interact with them, and follower details such as job title, location, and industry.

How much likes, comments and other responses matter to the person who posted. LinkedIn’s feed considers whether a response could encourage that person to keep posting, including people who get little feedback.

Personalized suggestions of creators to follow in My Network and the home feed, based on how well a creator’s profile matches a member’s interests. These suggestions can help members discover relevant creators even if they have few followers.

Members who create and share content on LinkedIn to build an audience and connect with their followers.

A technique described in a Microsoft patent for making a published post also appear in a relevant group's feed.

An EU law that requires online platforms such as LinkedIn to provide more transparency, user controls, and accountability. For LinkedIn, its requirements include publishing transparency reports, assessing risks arising from the service, and undergoing independent audits.

A system LinkedIn uses to detect and prevent discriminatory job listings across all job types, helping protect job seekers and support fairer hiring.

Feedback that tells LinkedIn you are not interested in certain content, such as hiding a post you don’t find relevant or useful. LinkedIn considers this feedback when choosing posts to recommend in your feed.

An earlier recommendation system at LinkedIn that trains on individual activity events and their details, using roughly the past 90 days of data. LinkedIn compares it with its newer Generative Recommender system.

The impact a viewer’s actions on a post can have on their network, such as a reshare causing their connections to see it in their feeds. LinkedIn’s Feed ranking can consider this wider impact alongside the viewer’s own likely response.

Measures of members’ delayed responses to the LinkedIn Feed, such as visits during which they interact with content.

People who may see content because earlier viewers shared or acted on it. They are described in a Microsoft patent about ranking posts in a social network feed.

An independent part of LinkedIn Ireland that puts measures in place and checks compliance with the Digital Services Act (DSA). Its head reports directly to LinkedIn’s DSA board committee.

How long a member spends viewing a post or other content on LinkedIn. LinkedIn uses this to help decide what to show in the Feed, including when people read without liking, commenting or sharing.

Ads personalized with a member’s profile photo, name, or job function. They appear to the right of the Feed on LinkedIn’s website, have an “Ad” label, and are not shown in the mobile app.

A technique described in a paper and tested on LinkedIn’s job matching platform for changing how a recommendation system selects initial results as priorities shift. It adjusts the weight given to different goals while the system runs, without retraining it.

A research technique for choosing initial recommendations by splitting each item’s numerical representation into parts for different goals. Tested on LinkedIn’s job matching, it lets the system adjust how much each goal matters without retraining.

Posts that try to draw clicks or comments with prompts such as “Comment ‘Yes’ if you agree.” LinkedIn considers engagement bait low-quality content and is working to show less of it in the Feed.

Groups of people who coordinate likes, comments and shares on LinkedIn posts to artificially boost their visibility. LinkedIn does not allow this activity.

A writing feature shown when composing a post or message. LinkedIn is replacing it with tools that proofread your words and improve clarity without changing your voice.

An approach to making LinkedIn’s AI systems understandable, so people can see how they work, spot problems, and check whether they are fair, safe, and respectful of privacy.

Accounts with phony profiles or profiles that impersonate another person. LinkedIn works to block them before they appear on the platform and remove those it detects as fake.

A section of your profile where you can highlight your newsletter so visitors can quickly find the latest edition and subscribe.

A stream of posts and other professional updates on your LinkedIn homepage, including content from your network, companies you follow, and sponsored posts. By default, posts are personalized and ordered based on your interests and activity.

Tools that let you shape what appears in your LinkedIn feed, including following, unfollowing or muting people and companies, hiding posts, and telling LinkedIn you’re not interested in similar content.

A system LinkedIn uses to rank Feed posts by reading, in order, the posts a member has seen and how they responded. It uses patterns in that history to predict their actions on new posts.

A function described in a Microsoft patent for modeling how feedback, such as likes and comments, may affect how much a creator posts. It suggests extra feedback may encourage more posting, especially for creators who receive little feedback.

A measure described in a Microsoft patent that estimates how a member’s likelihood of creating content changes when they receive more feedback, such as likes, shares or comments.

In Microsoft's patented feed-ranking methods, first-pass rankers quickly score possible feed items to create shortlists. A later ranking stage scores those shortlists more carefully to choose what appears in the feed.

A system that stores and retrieves content for LinkedIn’s Feed and makes the first selection of posts that could be shown.

A link in LinkedIn’s network between a member and something they follow, such as a hashtag or company.

A page that shows recommended sources to follow, helping you find members and companies whose updates you want in your feed.

How often a suggestion to follow a creator on LinkedIn leads to a member following that creator.

People who receive your posts, articles, and shares in their LinkedIn feed. They can follow you whether or not they’re connected to you.

A page where you can see and manage who you follow and who follows you on LinkedIn.

A technique described in a Microsoft patent that limits how often one person sees the same content item within a set period.

A technique LinkedIn’s researchers have described for spreading Feed posts through direct connections, follows and engagement by members connected to the viewer’s network.

A conversation in LinkedIn messaging where you can communicate with several connections in one place.

Forums that members join to discuss shared interests or experiences on LinkedIn. They are created and managed by individual members and can be public or private.

Groups of LinkedIn members used for long-term comparisons to help identify unwanted effects from changes to the systems that decide what people see.

A queue for people to review content, described in a Microsoft patent as one tool content classifiers may use to detect spam and limit the spread of low-quality, unprofessional content.

An option shown after you ignore a connection invitation that prevents the sender from inviting you again. It also gives LinkedIn feedback it may use to restrict the sender’s account.

In the cited patent, immediate viewers are people whose feeds show an item because someone publishes or shares content.

Address-book entries brought into a social network from phone or email contact lists. The cited Microsoft patent describes using uploaded contacts to suggest connections while protecting members’ privacy.

The percentage of times your posts appeared on screen to members who follow you or are connected to you.

A test program that lets invited creators earn money when video ads play before their eligible videos on LinkedIn.

Artificial interactions on LinkedIn, such as likes, comments, and shares, often created through automation or coordination to boost a post’s visibility.

A technique described in a LinkedIn patent for estimating whether profile details are inaccurate by comparing a member’s profile and activity with patterns from other members. It assigns confidence scores to details and adjusts search results when a detail is judged inaccurate.

Notices shown within LinkedIn that encourage you to report content its automated systems detect as potentially breaking its rules.

A feature LinkedIn is testing during sign-up that lets new members choose topics they care about so their Feed can show relevant content from day one.

Notifications about jobs based on search criteria you choose. Your job alerts also help LinkedIn personalize the jobs it recommends to you.

A section of LinkedIn where you can search for jobs, see personalized job recommendations, and apply on LinkedIn or through an employer’s website.

The time LinkedIn’s systems take to find and return results, such as posts for your Feed or search.

A technique LinkedIn’s researchers have described for personalizing recommendations by adapting a small part of a recommendation model to a member’s recent interactions. It creates a compact summary of that adaptation for the main model to use when tailoring recommendations.

Advertising products that let you promote content to selected audiences on LinkedIn, using details such as industry and job role to decide who sees it.

A set of software tools LinkedIn built and shared publicly to measure whether AI systems treat different groups of members fairly. It can check for bias in the data used to train systems and compare how well systems work for different groups.

A feature that lets people and organizations broadcast live video to their LinkedIn networks in real time.

An agreement that authorized representatives must accept before creating a LinkedIn Page for their organization.

A technique LinkedIn’s researchers have described for finding posts from outside your network to recommend in the Feed. It scans the full collection of posts and applies filters before ranking matches.

A technique LinkedIn’s researchers have described for building AI models that rank Feed posts, job recommendations and ads to help decide what members see.

A way LinkedIn uses large language models—AI systems that understand text—to help order Feed posts and search results by how well they match a member’s interests or search query.

A technique LinkedIn uses to adjust an existing language model for different tasks through small add-on components called adapters, instead of building a new model each time. It helps LinkedIn update models efficiently as content patterns change.

Actions members take on LinkedIn, such as viewing content and interacting with posts. LinkedIn uses patterns in these actions over time to help choose relevant posts for their Feed.

Interactions members have with content on LinkedIn, including reactions, comments, saves, sends, and reposts.

How often job seekers accept outreach messages from hirers on LinkedIn.

A way to contact certain LinkedIn members you aren’t connected to through a shared context, such as being in the same group or working at the same company.

A feature for sending direct messages to other LinkedIn members, individually or in group chats. You can also contact some people you aren’t connected to through message requests or paid InMail messages.

A measure of how often content flagged by LinkedIn’s moderation systems actually breaks its rules.

A feed setting that shows the newest posts first, only from people you follow. It does not use your interests or activity to sort the feed or make recommendations.

A Feed sorting option that orders posts based on your interests and activity. The Top posts view uses this personalized ordering.

An area of LinkedIn where you can find people to connect with, access connection invitations, manage who you follow, and view your followers.

A possible restriction described in a Microsoft patent that allows some content labeled low quality to be shown only within a network.

A chart in audience analytics that shows your follower growth over the date range you select.

A series of regularly published articles on LinkedIn, usually about a specific topic. Members can subscribe to receive push, in-app, and email notifications when you publish a new article in the series.

Alerts about LinkedIn activity, such as new messages, invitations and comments on your posts. They can appear inside LinkedIn, arrive as alerts on your device or come by email, and you can control which types you receive.

A feature that lets you indicate interest in job opportunities and express job preferences that help shape your job recommendations.

People you aren’t connected to and don’t follow on LinkedIn, or posts from those people.

Members with permission to manage an organization’s LinkedIn Page. They can accept collaboration invitations for the Page and delete comments on its posts; some admin roles also allow them to add or remove other admins.

A model described in Microsoft patents that predicts how likely a creator is to post in the future based on how much feedback, such as likes, comments and shares, they receive.

A feature on your My Network page that suggests LinkedIn members you may want to connect with. Suggestions use shared connections, similar profile information and experiences, and contacts you’ve imported from your email or phone.

A way of grouping numbers, such as engagement rates, into ranges based on how they compare with other values, so LinkedIn’s Feed ranking model can better understand their size.

A setting that lets members in the European Economic Area (EEA) turn off personalization of job recommendations, which otherwise use their job searches, job alerts, profile, and LinkedIn activity.

A feed setting that lets members in the U.S. adjust their preferences for political content.

A restriction on how many posts a member or LinkedIn Page can publish within a certain period.

Updates that members share in the LinkedIn feed, including text, images, videos, links, documents and polls.

The initial stage when a large language model reads and processes its input. LinkedIn’s ranking systems can use this stage to score search results without generating text.

A way of scoring search results in which a large language model reads the search, member information and each result once, then returns a score without generating text.

A subscription for a company’s LinkedIn Page that gives the company access to Page verification.

How well creators suggested to a member match that member’s professional background.

The rules for what content and behavior are allowed on LinkedIn, including bans on harassment, misinformation, and hate speech. Members agree to follow them when they join.

Details a member includes on their LinkedIn profile, such as their experience, company or school. LinkedIn uses this information to help suggest people to connect with and personalize job recommendations.

A tool LinkedIn uses to turn member and post information into text in a set format that an AI language model can process.

Personal attributes, such as age and gender, that advertisers must not use to discriminate on LinkedIn. LinkedIn also restricts whether advertisers can use these attributes to choose who sees an ad.

An option that lets LinkedIn Page super admins and content admins choose whether to publish an article or newsletter from their personal profile or a Page they manage.

The Microsoft patent describes using formulas to score content against goals, such as engagement or job relevance, to help decide its order.

How widely your content is seen on LinkedIn, including by people who follow or are connected to you and people outside that group. LinkedIn’s impression count estimates how many times a post appeared on screen.

Ways to seek further review of a decision about your LinkedIn content or profile, linked from the notice you receive after your appeal is resolved.

A score for a member’s reputation on a particular topic, described in LinkedIn patents. The patents describe how a higher score could make the member’s comments or articles appear more prominently.

Sharing someone else’s post or article with your own LinkedIn audience.

A feature that lets you transfer information from an uploaded resume to your LinkedIn profile.

Group sessions where LinkedIn’s content moderators address common review errors to help them apply content policies more consistently.

A tactic LinkedIn is testing in human review of harmful content that hides information about previous reviewers.

A pattern described in a Microsoft patent where already prominent contributors keep getting shown in a social network feed and gain more attention.

A way to find people, jobs, companies, posts, and more on LinkedIn using keywords or everyday language. You can narrow the results with categories and filters.

One-click shortcuts for narrowing job searches using related companies or job titles. They are part of a technique described in a Microsoft patent.

A feedback option that lets members flag feed posts or comments that feel generic or lack a clear perspective or substance. It helps LinkedIn improve content quality without starting a reporting process or indicating a policy violation.

A way for members to share demographic information, such as gender and disability, with LinkedIn. This information helps LinkedIn assess how different groups experience the platform and check for gender bias in how posts are distributed and ranked in the Feed.

The stage when LinkedIn’s live systems select or score recommendations for a member, using models that have already been trained.

In Microsoft patents about blending sponsored and organic items, such as job listings, a shadow bid is a hypothetical bid assigned to unpaid content so it can be compared with paid content when choosing positions in a list.

An option in LinkedIn’s article edit menu that gives you a link to share an unpublished article for feedback. Only people you share the link with can see the draft.

A number estimating how likely an organization’s members are to share a piece of content, used to rank content for sharing recommendations. Microsoft described this technique in a patent.

An interface for creating a LinkedIn post, where you can add collaborators before publishing.

A collection of exclusive video content made by creators and publishers around themes such as AI, small business, or female leadership that brands can sponsor on LinkedIn.

A LinkedIn page for a skill that you can follow.

A setting that lets you choose the order of feed updates or job search results, such as showing the newest feed posts first.

Markers placed around numerical values grouped by how they compare with other values, helping LinkedIn’s Feed model learn about quantities such as engagement rates.

A notice explaining why LinkedIn has moderated your content or restricted your profile. LinkedIn also publishes these explanations, without personal data, in the European Commission’s Digital Services Act database.

Rules that determine when LinkedIn restricts an account based on the type of violation and how many violations have occurred.

A control at the top of your profile that lets followers choose to receive a notification whenever you publish a new post on LinkedIn.

Posts LinkedIn recommends in your feed based on what is relevant to your professional identity, activity, and goals. They may come from people you are not directly connected with.

A LinkedIn Page administrator role that lets you choose which account to publish articles from and share Page updates directly from your homepage.

A form that lets LinkedIn members report media posted by another member that appears artificially created or misleading.

A way LinkedIn uses larger AI models to create training data for smaller, faster models. This helps the smaller models learn to identify new patterns of generic AI-generated content, known as AI slop.

A newsletter curated by LinkedIn’s News team that you can read for inspiration.

Subjects a member appears to care about, based on their professional background and how they interact with content. LinkedIn uses these interests to help choose relevant posts for their Feed.

How closely a creator’s topics match a member’s interests. LinkedIn uses this match to help suggest creators to follow based on what their profiles say they are about.

Collections of developing professional news and conversations tailored to your interests. LinkedIn’s editors and automated systems select and update related articles and posts to help you explore different perspectives and join the discussion.

A review LinkedIn carries out before launching new products or initiatives to identify and reduce risks involving privacy, security, safety, and legal compliance. It also includes checking algorithms for potential bias.

How many different people interacted with content. A Microsoft patent describes a way to estimate this using anonymous user groups.

These measurements capture people’s immediate reactions to the LinkedIn Feed, such as how often they interact with an update.

A check that confirms your identity, workplace, or educational institution on LinkedIn. Once completed, it adds a badge beside your name on your profile to help others assess your authenticity.

A video post shared directly on LinkedIn, either recorded in the app or uploaded from a video you recorded earlier.

Measures of how often someone interacts with videos in different ways. A Microsoft patent describes a technique that uses these rates from past video activity to predict whether someone will interact with a video and help choose videos when a content request comes in.

A tab in some LinkedIn mobile experiences where members swipe through short videos selected by editors. It shows videos from the Feed, with the same underlying posts, comments, and reactions.

A setting that lets you choose whether all LinkedIn members or only people you’re directly connected with can follow your updates.

A feature that explains why you were included in an ad’s intended audience, who the advertiser is, and who paid for the ad. You can open it from the ad’s More menu.