Un guide complet sur la terminologie des médias sociaux. De la modération et des analyses à l'engagement et aux concepts techniques, apprenez les termes clés que chaque responsable de marque doit connaître.
Agentic AI refers to AI systems that can autonomously plan and carry out multi-step tasks toward a goal with limited human input, rather than just responding to single prompts. In comment management, an agentic system can monitor, classify, hide, and reply to comments continuously on its own, within guardrails set by the brand.
Lire la définitionAn AI shopping assistant is a conversational AI tool, such as ChatGPT shopping, Perplexity, or Gemini, that helps people research and choose products by answering questions and making recommendations. These assistants read reviews, social signals, and comment sections, which makes a brand's online reputation a direct input to what they recommend.
Lire la définitionAPIs and webhooks are technical integration methods that allow different software systems to communicate. APIs enable on-demand data requests, while webhooks push real-time event notifications automatically.
Lire la définitionAstroturfing is the practice of creating fake grassroots engagement — making paid, coordinated, or bot-driven activity appear as organic public opinion.
Lire la définitionBrand safety refers to the strategies and tools used to protect a brand's reputation by ensuring its advertising and social presence do not appear alongside harmful, offensive, or inappropriate content.
Lire la définitionBrigading is a coordinated attack where a group of users flood a post, comment section, or account with negative comments, reports, or downvotes to overwhelm the target.
Lire la définitionComment farming is the practice of using bot accounts, engagement pods, or paid workers to generate large volumes of comments on social media posts to artificially inflate engagement metrics.
Lire la définitionComment filtering is the process of automatically screening and managing comments on social media posts using keywords, AI models, or rule-based systems to remove unwanted or harmful content.
Lire la définitionComment spam refers to unsolicited, irrelevant, or promotional messages posted in comment sections of social media posts, blogs, and other online content. It degrades user experience and can harm brand reputation.
Lire la définitionA community guidelines violation (CGV) occurs when a piece of content — a post, comment, or message — breaks a social media platform's published rules for acceptable behavior on the platform.
Lire la définitionCommunity management is the practice of building, growing, and nurturing an engaged online audience around a brand through active moderation, conversation facilitation, and relationship building on social media.
Lire la définitionContent moderation is the process of reviewing and filtering user-generated content to ensure it complies with platform guidelines, community standards, and brand safety requirements.
Lire la définitionCoordinated Inauthentic Behavior (CIB) is Meta's term for networks of fake accounts working together to manipulate public discourse, deceive people about who is behind the activity, or artificially amplify content.
Lire la définitionCopypasta is text that is copied and pasted across many comment sections, posts, or messages — often used for spam, trolling, or viral memes that overwhelm comment sections.
Lire la définitionThe Digital Services Act (DSA) is a European Union regulation, fully applicable since 17 February 2024, that sets binding rules for how online platforms handle illegal and harmful content, including user comments. It requires notice-and-action systems, clear statements of reasons for content removals, and transparency reporting, with fines of up to 6% of global annual turnover.
Lire la définitionA dog whistle is coded language in comments that appears innocent to most readers but carries a hidden hateful, political, or discriminatory meaning understood by a specific in-group.
Lire la définitionDoxxing is the act of publicly revealing someone's private personal information (real name, address, phone number, workplace) online without their consent, typically to harass or intimidate.
Lire la définitionEngagement bait is content or comments deliberately designed to provoke reactions, replies, or shares by exploiting emotional triggers — typically controversy, outrage, or curiosity.
Lire la définitionEngagement rate is a social media metric that measures the level of interaction a piece of content receives relative to audience size, expressed as a percentage of likes, comments, shares, and saves.
Lire la définitionFake followers are bot-generated or purchased accounts that follow a social media profile to artificially inflate follower count without providing genuine engagement.
Lire la définitionGenerative Engine Optimization (GEO) is the practice of structuring content so that AI answer engines such as ChatGPT, Perplexity, Google AI Overviews, and Gemini cite and surface it in their responses. Where SEO targets ranked links, GEO targets being quoted directly inside AI-generated answers.
Lire la définitionA hate raid is a coordinated mass attack on a live stream or social media account where large numbers of bot or troll accounts flood the target with hateful messages, slurs, and harassment.
Lire la définitionHidden Words is a native Meta feature on Instagram and Facebook that automatically hides comments and message requests containing words, phrases, or emojis a user has chosen to filter. It is a free, built-in keyword filter, but it only matches exact terms and cannot understand context the way AI moderation can.
Lire la définitionImpersonation is when someone creates a fake social media account that mimics a real person, brand, or organization to deceive followers or steal engagement.
Lire la définitionLLM moderation is the use of large language models to classify and act on user-generated content by understanding meaning and context, rather than matching keywords. It detects spam, scams, abuse, and sentiment in comments even when they are disguised, misspelled, or implied, which keyword filters cannot do.
Lire la définitionNotice-and-action is a content-moderation mechanism, mandated by the EU Digital Services Act, that lets users flag illegal content and obligates the hosting service to review the notice and act on it in a timely, non-arbitrary way. It is the core process platforms must offer for reporting and removing illegal comments.
Lire la définitionThe Online Safety Act is a United Kingdom law, regulated by Ofcom, that requires services hosting user-generated content to protect users from illegal content and to keep children safe online. It imposes duties to assess and mitigate risks, with fines of up to £18 million or 10% of global annual revenue for non-compliance.
Lire la définitionA restricted word list is a curated set of words and phrases that triggers automatic moderation actions when detected in social media comments, helping brands filter offensive, spammy, or harmful content.
Lire la définitionSentiment analysis is the use of natural language processing and machine learning to identify and categorize the emotional tone behind text, determining whether content is positive, negative, or neutral.
Lire la définitionShadow banning is the practice of silently restricting a user's content visibility without notifying them. The user can still post, but their content is hidden from other users or deprioritized by the platform's algorithm.
Lire la définitionSocial listening is the practice of monitoring social media platforms for mentions of your brand, competitors, and industry topics to gain actionable insights about audience sentiment and market trends.
Lire la définitionA social media bot is an automated program that operates on social platforms to perform tasks like posting content, liking posts, following accounts, or generating comments without direct human involvement.
Lire la définitionA social media troll is a person who deliberately posts inflammatory, offensive, or disruptive content online to provoke emotional reactions and derail constructive conversations.
Lire la définitionSocial proof is the psychological phenomenon where people look to the actions and opinions of others to guide their own decisions, widely leveraged in marketing through reviews, testimonials, and engagement metrics.
Lire la définitionA sock puppet account is a fake online identity created by someone who already has an account, used to deceive others by pretending to be a different person — often to support their own views, attack opponents, or evade bans.
Lire la définitionA statement of reasons is a clear, specific explanation that an online platform must give to a user when it restricts, removes, or demotes their content, as required by Article 17 of the EU Digital Services Act. It tells the affected user what action was taken, why, and how to appeal.
Lire la définitionA trusted flagger is an entity granted special status under Article 22 of the EU Digital Services Act whose notices about illegal content must be treated with priority by online platforms. Trusted flaggers are awarded the status by national regulators based on proven expertise and independence.
Lire la définitionUser-generated content (UGC) is any content such as text, images, videos, or reviews created by users rather than brands. It serves as authentic social proof and drives engagement on social media platforms.
Lire la définitionMettez ces concepts en pratique. Commencez votre essai gratuit et protégez les commentaires de votre marque sur les réseaux sociaux.
7-day free trial
