The Like Began as a Comment Shortcut—Then It Became a Ranking Signal

The Like button began as a response to repetitive, low-value comments, not as a documented attempt to rescue Facebook’s servers. What looked like a small interface shortcut became far more consequential once each click could measure interest and help determine which posts received attention.

That distinction still matters for creators. Meta’s published account of Feed ranking says a person’s previous likes inform predictions about what they may engage with, while the expected value of a like can carry more or less weight than a comment for different users; Meta’s ranking explanation also makes clear that likes are only one input in a multilayered system.

The original problem was noisy conversation

In 2007, acknowledging a Facebook post generally meant writing a comment. Popular posts consequently accumulated strings of brief responses such as congratulations or simple expressions of approval. Those messages were legitimate interactions, but they made substantive comments harder to find and did not fit the cleaner product Facebook wanted to build.

Product manager Leah Pearlman and colleagues explored a universal, low-effort response under the internal name “Props.” According to Fast Company’s account based on interviews with the team, Facebook had about 30 million users in the summer of 2007, and an internal group produced an early “awesome button” prototype that July. The proposal drew interest from several teams because the same action could acknowledge a post, support Feed decisions and reveal preferences.

This history does not support the neat story that a tiny database record was invented primarily to reduce infrastructure costs. A Like still had to be stored and associated with both a person and a post. The documented product objective was to make approval easier and to prevent one- or two-word affirmations from overwhelming the comments that carried more information.

No single company invented online approval

Facebook made the Like globally recognizable, but approval mechanisms were already circulating across consumer websites. Yelp, Digg, YouTube, Vimeo and other services had experimented with lightweight ways to rate or respond to user contributions. The modern button emerged from that broader product culture rather than from one isolated flash of invention.

FriendFeed introduced its own Like in October 2007, while Facebook’s version did not launch until February 9, 2009. A 2025 Associated Press reconstruction describes the feature as a collective innovation with several precursors and notes that Facebook spent nearly two years debating its implementation. Facebook had already launched the button when it agreed to acquire FriendFeed later in 2009.

The chronology makes two simplistic claims difficult to sustain. Facebook did not originate the general idea of one-click approval, and FriendFeed did not simply hand Facebook a finished feature before Facebook began thinking about the problem. Similar designs developed across several companies, while Facebook’s distinctive achievement was attaching the interaction to a network large enough to establish a common convention.

A shortcut became a measurement system

A comment can contain praise, disagreement, a question or a joke. A Like compresses that ambiguity into a standardized event that can be counted and compared. That loss of nuance is precisely what makes the response convenient for users and tractable for software.

Once visible counts appeared beside posts, the button served three audiences at once. Users received a low-friction way to acknowledge someone; creators received an immediate popularity indicator; and the platform received structured behavioral data. A feature intended to tidy conversation therefore became part of the machinery that distributed conversation.

Facebook later expanded the interaction rather than abandoning it. Reactions added several emotional responses around the original Like, but they preserved the central design: a person selects a standardized response, the platform records it, and other people may see an aggregate count. The vocabulary became richer without returning to the open-ended detail of a comment.

What a Like tells creators—and what it does not

For a creator, a Like is evidence of an action, not a complete diagnosis of audience sentiment. It can indicate recognition, support, agreement, habit or simply a desire to acknowledge a post without composing a reply. The interface does not reveal which of those motives produced the click.

Nor does a high Like count translate mechanically into equivalent reach. Feed systems consider many predicted actions and contextual signals, and their weights can differ by person. A post that earns quick approval may perform differently from one that prompts comments, viewing time, sharing or other behaviors, even when the visible counters appear similar.

The practical consequence is to treat Likes as one comparable signal within a defined context. They are useful for comparing similar posts shown to similar audiences over a similar period, but less reliable when formats, distribution or audience size differ. The number is precise; its meaning remains conditional.

The real scaling breakthrough was social, not computational

The Like did help Facebook accommodate interaction at enormous scale, but primarily by lowering the effort required from people. Millions of users could respond without finding words, and the resulting activity arrived in a standardized form that a Feed could process. That is a product and behavioral form of scaling, not proof that server strain caused the feature to be built.

Its broader legacy follows from that trade-off. The button made participation easier, supplied creators with visible feedback and gave platforms a repeatable signal of interest. At the same time, it reduced many different human reactions to a small set of countable choices.

The Like conquered the internet because it solved several problems after launch, not because it began as an infrastructure trick. It cleaned up comment threads, encouraged lightweight participation and eventually became an input to personalized distribution. The humble shortcut endured because platforms could turn a moment of approval into both public feedback and machine-readable evidence.

Also read:

• The AI Scientist Hits Nature: Scaling Scientific Discovery Like Code
• SaaS Google Ads Can Optimize the Wrong Lead: Fix the Signal Before Scaling
• How Tokyo Became an Unexpected Haven for China’s Middle Class