Social networks are no longer networks of people. They are networks of recommendation. TikTok, Instagram, Facebook and YouTube now show content chosen by an algorithm. We see only a small fraction of what the people we follow post. Most of our feeds are filled with accounts we have never heard of. This shift fundamentally changes how reach and engagement work, and what a follower’s role is.
The clearest example is TikTok. The Washington Post recently analysed data from more than 1,100 volunteers who gave the newsroom the full watch history of their TikTok accounts for six months. That came to roughly 14.8 million videos watched. According to this analysis, users saw content from creators they follow in only 10% of their time on the app. The rest was videos recommended by the algorithm.
On Instagram and Facebook the shift is clear in the numbers. According to the Socialinsider report “Social Media Reach: Statistics for 2025”, Instagram’s average organic reach is 3.5% of followers, down 12% year on year, while Facebook sits at around 1.2%. Socialinsider states: “The average reach rate for Instagram is 3.50%, marking a 12% decrease compared to 2024. Facebook’s average reach rate stands at 1.20%.” Meta itself admits that connecting people is no longer the main function of social networks.
Social.plus summarised this data as follows: “In 2025, the average Instagram post reaches just 3.5% of a brand’s followers. On Facebook, it is even lower, averaging around 1.65%.” The average post therefore reaches only a very small part of the audience, however large the creator’s or brand’s account is.
The trend shows up in engagement too. According to the “2025 Social Media Industry Benchmark Report” from RivalIQ, average engagement rates across the main platforms fell year on year: Facebook by 36%, Instagram by 16%, TikTok by 34% and X by 48%. RivalIQ sums up the results this way: “Engagement rates continued to decline across most major platforms in 2025, with Facebook down 36%, Instagram down 16%, TikTok down 34%, and X down 48% year-over-year.”
A similar pattern shows up on YouTube Shorts. According to Zebracat’s analysis (2025), a large share of Shorts views come from users who do not subscribe to the channel. The analysis states that “74% of all Shorts views come from non-subscribers, while only 26% come from subscribers”. Shorts therefore work mainly as a recommendation feed, not as a channel built on a subscriber base.
The data from TikTok, Meta and YouTube confirm each other
Taken together, the data confirm a fundamental shift: social networks no longer work on the logic of “I follow, I see”. They work on the logic of “the algorithm chooses, I see”. Followers remain an important signal of trust and reputation, but they no longer act as a distribution unit. Recommendation models now decide distribution, and they are guided not by social relationships but by the likelihood that content holds attention. Creators who want a stable audience therefore have to build it outside algorithmic environments, in newsletters, podcasts, paid communities or on their own platforms.
“The FYP ate the follower” is not a criticism but a description of the new reality. The feed works like a television run by AI, and followers no longer decide what people see. All the major platforms show the same trend. Content wins when it holds attention, not when the creator has a large base of subscribers. Creators who can combine algorithmic formats with stable distribution outside the feeds have a better chance of long-term growth in this change. That is also why podcast popularity is growing. People are looking for a space where they are not pushed by a recommendation model, and where they feel real contact with a person rather than with an anonymous calculation.
The impact on brands and their collaborations
All of this has a direct impact on brand collaborations with creators. The number of followers no longer tells you who a creator actually reaches. In an algorithm-driven environment, what matters more is how a creator performs in the real feed, how consistently they can hold attention, and whether their work fits naturally into the recommendation model. A creator with a smaller audience who understands how the platform works often delivers a bigger effect than someone with a million subscribers whose content does not perform in the algorithm.
The value of creators who have their own distribution channels outside social networks is also rising. Newsletters, podcasts, communities or other platforms under their control give a collaboration a stability that algorithmic feeds no longer offer. At the same time, content has to be native and natural. Recommendation models favour work that looks like an organic part of the profile, not an advertisement inserted from outside.
Social networks have changed the rules of the game. Reach no longer comes from the number of followers. It comes from how well a format fits the algorithm and from a creator’s ability to build their own distribution network outside the platforms. For brands and creators alike, this means one thing. Success is now measured differently than before.
Comments