What Happens To A Brand When The Algorithm Changes

Every social-first brand is a tenant. The audience it has spent years assembling sits on infrastructure it does not control, governed by rules it is not told in advance.
This is understood in the abstract and almost never planned for. Teams build reporting, targets and headcount around a level of organic reach that is provided at the platform's discretion and withdrawn the same way.
What actually happens when the algorithm changes
Reach falls before anyone knows why. The first response is usually to blame the content, so the team changes the creative, which does nothing, because the variable that moved was not on their side of the relationship.
By the time the change is understood, weeks of production have been spent optimising against a system that had already moved. This is the real cost: not the lost reach, but the effort misdirected while diagnosing it.
Why brands over-index on whatever is currently working
Because it works, and because performance reporting rewards doing more of it. A format delivering unusual reach attracts budget and headcount, and within two quarters the operation is structurally dependent on one format on one platform.
That concentration is invisible while the numbers are good. It is only legible as risk afterwards, which is why it is almost never argued against at the time.
What can actually be owned
Very little, and it is worth being precise about which little. An email list is owned. A phone number with consent is owned. A returning direct-traffic audience is owned. A community on a platform is not, however engaged it is.
Followers are the clearest example of a metric that feels like an asset and behaves like a lease. A brand with a large following and no way to reach those people off-platform has an audience it cannot contact.
The realistic hedge
Not abandoning platforms, which would be silly, but making sure every platform relationship has an off-ramp. Something the audience can do that moves them somewhere the brand controls, offered continuously rather than in a panic.
The conversion rate on that will be low. It is supposed to be low; it is insurance, and the point is that the small proportion who take it are the part of the audience that survives a change in the terms.
How to notice earlier
Track reach separately from engagement, and track it as a ratio to followers rather than as an absolute. A sustained fall in that ratio across formats is a platform signal, while a fall in one format is a content signal, and conflating them is what produces the wasted weeks.
Most teams do not separate these because the dashboard does not, and the dashboard does not because the platform supplying it has no interest in making the distinction easy to see.
