Comments as a Feed Ranking Signal
This page summarizes evidence showing that Facebook’s Feed system treats comments as an important ranking and optimization signal. Facebook’s business model relies on on-platform time. Because comments are the primary driver of return visits (via notifications) and long-session durations, the algorithm is intentionally designed to "starve" content that does not contribute to these goals. Turning off comments is essentially telling the algorithm: "This post is a dead end for user session time," and the AI responds by showing it to fewer people. While Meta does not appear to publish a rule saying that turning comments off directly “punishes” a post, the available evidence strongly suggests that disabling comments removes one of the platform’s core engagement signals and may reduce distribution as a result.
Meta says comments are a ranking signal
Meta has said that Feed ranking considers interaction signals such as reactions, comments, and shares when deciding how prominently a post appears.
- Comment activity is one of the signals that can move content up or down in distribution.
- Meta has also emphasized “meaningful interactions,” including back-and-forth discussion in comments.
- Pages that fail to generate reactions or comments may see declines in distribution.
Meta says Feed ranking predicts whether users will comment
Meta has explained that its systems make predictions about what users are likely to do with a post, including whether they are likely to comment on it.
- Comments are not merely counted after a post is published.
- They are part of the predictive ranking logic used to score content.
- If comments are disabled, the post cannot produce comment outcomes in the usual way.
Meta engineering materials treat comments as a core optimization target
Meta’s engineering explanations of News Feed ranking describe models that score possible outcomes such as likes, comments, and shares, then combine those values into an overall ranking score.
- The system evaluates multiple forms of engagement.
- Comments are treated as one of the meaningful actions the platform seeks to predict and optimize.
- This supports the view that comments are central to Feed ranking rather than incidental.
Independent research links comments to organic reach
A peer-reviewed study analyzing 1,025 unpaid Facebook Page posts found that organic reach was positively correlated with interactions, including comments.
- The findings are consistent with the hypothesis that comments help distribution.
- The study supports the idea that comment activity aligns with algorithmic advantage.
- This does not prove an explicit penalty for disabling comments, but it supports the practical effect.
Broader research supports the importance of highly commented posts
Research on social media ranking and platform incentives indicates that comments have long been one of the interaction types associated with stronger performance. Facebook’s broader algorithmic evolution has increasingly emphasized social interaction and meaningful engagement.
- Historically, ranking systems considered clicks, likes, comments, and time spent.
- Later changes more heavily emphasized meaningful social interactions.
- Highly commented posts can receive more visibility because they generate return visits, notifications, and longer on-platform sessions.
- Disabling comments removes one of the platform’s strongest signals for conversation and re-engagement.
References
- Lada, A., Wang, M., & Yan, T. (2021, January 26). How machine learning powers Facebook’s News Feed ranking algorithm. Engineering at Meta.
- Metzler, H., & Garcia, D. (2024). Social drivers and algorithmic mechanisms on digital media. Perspectives on Psychological Science, 19(5), 735–748.
- Mosseri, A. (2018, January 11). Bringing people closer together. Meta.
- Mosseri, A. (2018, May 22). News feed ranking in three minutes flat. Meta.
- Pócs, D., Adamovits, O., Watti, J., Kovács, R., & Kelemen, O. (2021). Facebook users’ interactions, organic reach, and engagement in a smoking cessation intervention: Content analysis. Journal of Medical Internet Research, 23(6), e27853.