Engagement bait is content engineered to farm interactions instead of earning them: tag-a-friend demands, vote-with-a-like polls, planted mistakes fishing for corrections, outrage manufactured for the comment storm. Platforms name it in their policies and demote what they detect.
Bait works on the raw counters, which is exactly the problem: it harvests engagement that carries no real interest signal, polluting the data feeds rank on.
Meta has published policy against baiting patterns since 2017, and detection demotes baited content in ranking rather than banning accounts. The penalty is quiet: your counters look fine while reach decays.
The line runs through content: a genuine question invites thought; bait rents reflexes. "What would you add to this list?" is a call to action. "Comment YES if you agree" is bait.
The long-term cost is worse than the demotion: a baited audience learns to interact emptily, and empty interaction trains the algorithm to misclassify your account. You're poisoning your own signal.
Not if the comment has content. Questions with real answers are legitimate; demanding token replies (YES, emoji votes) is the baiting pattern.
Rarely banned, routinely demoted: platforms suppress the content's distribution. Repeat baiting suppresses the account's overall reach.
Detection is imperfect and some formats ride the line. The trade is still bad: the accounts behind them buy short spikes with long-term reach decay and audiences trained to reflex.
Related terms
Every definition lives in the full glossary, and the services behind the vocabulary are on the services page.