Trolling
Also known as: internet troll, troll account, don't feed the trolls, concern trolling, sea-lioning
Trolling describes the practice of posting deliberately provocative, inflammatory, misleading, or offensive content in online spaces with the intent of provoking emotional reactions (anger, distress, confusion) in others.
The term originates from fishing ('trolling' for fish, slowly dragging a baited line to attract bites) adapted to baiting people online. Trolling spans a wide spectrum: from light-hearted pranks and impish humor within community norms to severe targeted harassment campaigns. The term has been diluted through common usage: many people describe anyone who disagrees with them online as 'trolling'.
Trolling, explained at three levels
At home
A troll is someone who posts comments or content online specifically to upset people, start arguments, or get attention, not because they have a genuine point to make, but because they enjoy the reaction. The classic advice is 'don't feed the trolls': if you respond to a troll (even to correct them or criticize them), you give them exactly what they wanted: attention and emotional reaction. Trolling works because it is hard to ignore content that is personally offensive or factually wrong, so many people respond and inadvertently reward the troll's behaviour.
At work
Organizations with public social media presences encounter trolling regularly: negative, provocative, or false comments on posts, reviews, or community spaces. The professional response depends on troll type: spam and bots (remove and block), legitimate criticism framed aggressively (respond professionally, address the substance), coordinated attack campaigns (document, escalate to platform trust and safety, consider statement). Distinguishing genuine dissatisfied customers from pure trolls is important: treating legitimate concerns as trolling damages customer relationships; treating trolls as legitimate concerns wastes resources.
At home
Trolling types on a spectrum: classical trolling (humor-based provocation, often within community norms), concern trolling (pretending to agree with a community's values while undermining them with bad-faith questions), sea-lioning (following someone with persistent, seemingly polite questions designed to exhaust and frustrate), and coordinated harassment (organized groups targeting individuals for abuse). Platform community managers and researchers use these distinctions to identify different risk levels: classical trolling is usually low-harm; coordinated harassment using trolling mechanics can cause serious psychological harm to targets.
At work
Brand community management for trolling: most community guidelines distinguish prohibited content (threats, personal attacks, hate speech) from protected content (legitimate criticism, negative opinions, disagreement). Banning users simply for criticizing an organization carries legal risk mainly for governments and public officials, whose social accounts can count as public forums in some jurisdictions; for private companies the bigger risk is reputational. Community managers should apply consistent standards to all users (not banning only unfavourable opinions), document moderation decisions, and have escalation paths for coordinated attacks that exceed normal moderation capacity.
At home
The psychology of trolling: research (Buckels et al., 2014, published in Personality and Individual Differences) found that self-identified internet trolls scored higher on the 'dark tetrad' traits (psychopathy, Machiavellianism, narcissism, sadism), with the strongest correlation being sadism: enjoyment of others' distress. More recent research distinguishes 'dark' trolling (motivated by genuine desire to harm) from 'joking' trolling (provocation within social norms for entertainment). Platform design matters: anonymity increases trolling behaviour; accountability mechanisms (real names, reputation systems) reduce it. The internet's default of anonymity in comment sections structurally enables trolling.
At work
State-sponsored trolling operations represent the professionalization of trolling for political ends: the Russian Internet Research Agency (IRA), documented in the Mueller Report and subsequent research, employed thousands of people to create fake accounts, amplify divisive content, and coordinate trolling campaigns targeting political and social fault lines in the US and other democracies. The IRA's operations were designed to look like organic grassroots American voices and exploited the same psychological mechanisms as individual trolling at industrial scale. This form of coordinated inauthentic behaviour has prompted platform transparency reports, election integrity teams, and disclosure requirements for political advertising.
Real-world examples
At home
Nadia moderated an online community for a specialty hobby. A recurring user posted provocatively in threads about controversial topics, not to contribute to discussion, but to see members argue with each other. She recognized the pattern as classical trolling (amusing themselves with the reactions rather than engaging genuinely) and implemented a 'boring the troll' strategy: the community adopted a norm of responding to obvious provocations with a brief, neutral acknowledgment and moving on, depriving the troll of the emotional reactions they sought. The user lost interest within a week.
At work
Thomas ran social media for a consumer brand during a product quality controversy. A coordinated campaign of negative comments flooded their Instagram: many from accounts created within the past 30 days, all using similar language patterns. He identified this as organized trolling rather than genuine customer dissatisfaction, documented the pattern for the platform's trust and safety team, and applied a two-track response: flagging fake accounts to Instagram for removal while responding substantively to real customer concerns about the quality issue. Treating the coordinated campaign as genuine customer feedback would have rewarded the campaign's goals.
Common confusions
- Many people describe anyone who disagrees with them online as 'trolling' — Trolling has a specific meaning: provoking emotional reactions without genuine intent to engage on the substance. Disagreement, criticism, negative opinions (even harshly expressed), and persistent challenges to someone's position are not trolling when the person holds the view genuinely. Mislabelling all criticism as trolling is itself a dismissal tactic that prevents engaging with legitimate challenges. The distinguishing test is intent: is the person trying to make a real point, or purely seeking an emotional reaction? Bad-faith interaction can still hold a genuine position.
- Some platforms and users believe that removing trolls' content and banning them is always the most effective response — Platform removal is necessary for severe harassment but can sometimes amplify attention to the troll's original message through the 'Streisand effect' (publicizing an attempt to suppress something). For low-level trolling, ignoring (removing emotional reactions that reward the troll) is often more effective than visible removal. Research on content moderation shows that consistent, well-communicated community standards applied consistently over time are more effective at reducing trolling than reactive removal without clear rules: trolls adapt to moderation patterns and test boundaries.
For analysts
When helping clients develop social media community management guidelines, the trolling framework is a core component. Recommendations: define categories of content with different response protocols (prohibited removal, concern trolling neutral response, coordinated attack escalation), establish clear community guidelines publicly so removal decisions can reference a known rule rather than appearing arbitrary, train community managers on distinguishing genuine criticism from bad-faith provocation, and document patterns of coordinated attacks for platform reporting. The goal is a community environment where genuine participants feel safe and trolls find no reward for their behaviour.
History & origin
The term emerged in early internet communities (Usenet, early forums, IRC) in the 1990s to describe users who posted provocative content to generate reactions. Early trolling had a mischievous, prankish quality: posting absurd claims to watch others argue seriously, asking obviously inflammatory questions in specialist communities to provoke experts. As platforms scaled, trolling professionalized: coordinated campaigns targeting specific individuals or groups, politically motivated trolling farms (documented in Russian Internet Research Agency activities in the 2016 US election), and harassment campaigns using trolling tactics. The 'golden rule of the internet' ('don't feed the trolls': ignoring trolls deprives them of the emotional reactions they seek) became common advice. Platform moderation evolved to include troll detection and account removal. The line between trolling and targeted harassment, political disinformation campaigns, and organized pile-ons has blurred significantly.