Every marketer over a certain age has the story about when a post to a page reached most of the followers. The standard explanation for what happened is that platforms restricted reach to force advertising spend.
That is true in part and it is not the main thing that happened, and understanding the rest changes what you do about it.
The supply problem
The mechanical explanation and the largest factor.
Feeds are finite. A user sees a limited number of items in a session, determined by how long they scroll.
The number of accounts producing content, and the volume each produces, has risen enormously. Every business, every publisher, every creator, posting more frequently.
So the amount of content eligible for a given feed grew by orders of magnitude while the space available stayed roughly constant.
The consequence is arithmetic rather than policy. If eligible content grows tenfold and slots stay the same, average reach falls by ninety percent. No decision by anybody is required.
The ranking change
The second factor and the one with the most strategic consequence.
Feeds moved from chronological to ranked, and ranking optimises for engagement — whatever keeps people using the product.
Which means content competes on predicted engagement rather than on recency or on whether somebody followed the account.
Following an account became a weak signal rather than a subscription. The platform interprets it as evidence of interest, weighs it alongside everything else, and shows you what it predicts you will engage with.
This is why a page with a hundred thousand followers can reach two thousand of them. The followers exist; the ranking does not consider them entitled to anything.
The shift away from follow graphs entirely
The most recent development and the one that has changed the game most.
Several major platforms now surface a large proportion of content from accounts the user does not follow, selected by predicted interest.
That cuts both ways for a business. Follower count matters less than it did, which is bad news for anybody who invested in building one.
And distribution is available to accounts with no audience at all, if the content performs, which is genuinely new and is why small accounts can now reach large numbers unpredictably.
The practical effect is that each piece of content is assessed largely on its own merits rather than inheriting an audience from the account.
What this means for how to work
Several things follow that are different from the old approach.
Follower count is a vanity number. What matters is reach and engagement per post, and the relationship between them and the follower count is now weak.
Posting frequently to stay visible works less well than it did, because volume without engagement trains the ranking against you.
The first few minutes matter enormously, because early engagement is a strong input to whether something gets distributed further.
And format compliance matters. Platforms distribute the formats they are currently promoting, and those change. This is the least satisfying finding and it is consistently true.
What is worth doing instead
The conclusion a lot of people have reached and it took us too long.
Build audiences on channels you control. Email principally, and anywhere else where reaching people does not depend on somebody else's ranking decision.
Use social platforms for discovery, which is what they are now good at, rather than as a distribution list, which they no longer are.
Accept that individual post performance is high variance and stop reading anything into a single result.
And measure what social actually delivers — traffic, sign-ups, sales — rather than engagement on the platform, which is the platform's metric rather than yours.
The honest position on paid
Paid social works and it is not a replacement for the old organic reach.
It reaches people who have not chosen to hear from you, at a cost, with the response rate that implies. Old organic reach was reaching people who had opted in, for free.
Treating paid as the way to restore what was lost leads to disappointment. Treating it as a distinct acquisition channel with its own economics is the workable framing, and it is a considerably less pleasant business than the one that existed a decade ago.
What still works organically
To avoid ending on pure pessimism, some things do.
Content that people send to each other privately, which is a large and largely invisible distribution channel that the metrics do not capture well.
Employee and founder accounts, which frequently reach further than brand accounts because the ranking treats individuals differently and because people engage with people.
Genuine community spaces, where the mechanics are different from a broadcast feed.
And replying, which reaches the person you reply to and everybody watching that conversation, at no cost and with none of the ranking disadvantage that a post from a brand account carries.