Why Follower Count Is a Weak Metric

Follower count is the number every small business owner checks and the number that explains the least. Here is the evidence, most of it published by the platforms themselves.

The distribution has moved off the follow graph

Meta publishes a quarterly Widely Viewed Content Report. For the fourth quarter of 2025, it reports that 46.8% of US Feed views "came from posts shared by people's friends, from Groups people had joined, or from Public Followers," while "41.0% came from in Feed recommendations, which show people content from sources they are not connected to."

Four in ten content views now come from accounts the viewer never chose to follow.

The direction is documented too. In July 2022 Mark Zuckerberg told investors that "about 15% of content in a person's Facebook feed and a little more than that of their Instagram feed is recommended by our AI from people, groups, or accounts that you don't follow." The two figures are measured differently, one a share of content and one a share of US views, so do not treat the gap as an exact trend. The shift is not in question.

YouTube said the same thing in different words back in 2021: "Recommendations drive a significant amount of the overall viewership on YouTube, even more than channel subscriptions or search."

Subscribing to something used to be how you saw it. Now it is one input among many.

On TikTok, followers are not even an input

TikTok's explainer is blunt: "While a video is likely to receive more views if posted by an account that has more followers, by virtue of that account having built up a larger follower base, neither follower count nor whether the account has had previous high-performing videos are direct factors in the recommendation system."

Followers there are a mailing list effect, not a ranking effect.

Some systems actively help small accounts

X publishes its For You ranking code. Among the adjustments applied after scoring: a "New-Author Boost: posts from authors whose impressions are below a threshold are lifted toward a target position."

There is also an "Out-of-Network Discount" that reduces posts from accounts the viewer does not follow, so followers still matter on X. But a small account is not simply at the back of the queue. There is documented machinery designed to give it a chance.

The number goes up and the effect goes down

Socialinsider's 2026 Instagram benchmarks, based on "35M Instagram posts, collected from 447,613 pages" active from January to December 2025, report average audience growth rates by account size: 22.00% for accounts with 1,000 to 5,000 followers, 20.29% for 5,000 to 10,000, 17.20% for 10,000 to 50,000, 13.62% for 50,000 to 100,000, and 11.25% for 100,000 to 1 million.

That is growth rate, not engagement rate, and it is one company's sample. It does still describe something real: the follower curve flattens as it grows, and the incremental follower is worth less than the one before it.

Hootsuite, reporting an analysis with data science agency Critical Truth of "over 1 million social posts across industries and social networks," puts the qualitative version plainly: "In general, you can expect your engagement rate to go down as your follower count goes up."

What followers still do

This is not an argument that followers are worthless.

They are a floor. Meta's own breakdown still shows nearly half of US Feed views coming from friends, Groups and followers. Those views are the ones you can somewhat rely on.

They are early evidence. Ranking systems need behavioral signal before they can predict anything. Your existing audience is where that first signal comes from, on every platform.

They are a list you do not own. Which is the real caveat. A follower is a permission the platform can reprice at any time. An email address is not.

What to watch instead

Sends and saves. X's published weights put share via copy link at 20.0 and share via DM at 5.0, against 0.5 for a like. Meta told investors its improved Reels ranking showed "particular strength in reshares and time spent, which are both strong indicators of better content-to-user matching." Content that gets forwarded privately is the strongest available signal that it landed.

Replies from people who know you. X applies a bidirectional follow reply boost of 15.0 on top of a reply's base weight of 5.0. A conversation with people who mutually follow you is worth more, in the model's own arithmetic, than a pile of passive likes.

Completion and dwell. YouTube ranks on watch time and satisfaction surveys. LinkedIn models the probability that you get skipped. TikTok weights finishing a video heavily. These are the metrics closest to what the systems actually optimize.

Repeat viewers. On any platform, the number of people who see several of your posts matters more than the number who saw one.

Business outcomes. Inbound enquiries, quote requests, bookings, people mentioning a specific post when they contact you. Boring, unfashionable, decisive.

The uncomfortable practical implication

If you have 400 followers and 30 of them reliably read, respond, and forward your posts, you have a better distribution asset than someone with 40,000 passive followers. The ranking systems are built to read exactly that difference.

It also means growth is not primarily a numbers game. It is a question of whether a small number of the right people find your posts worth reading, and whether you keep giving them something to read. That is a relationship problem, and relationships get built one specific person at a time.

When this was last confirmed

Meta's Widely Viewed Content Report figures are for the fourth quarter of 2025, published March 2026. TikTok's statement is from June 2020 and remains its most recent public position. X's ranking code was read in August 2026. The Socialinsider and Hootsuite figures are third-party samples, dated 2025 to 2026, and should be treated as descriptions of their samples rather than universal rates.