The short answer
An algorithm update is a change a platform makes to how its ranking system evaluates and distributes content, shifting which signals matter more or less. Updates happen continuously and without announcement. Only a small share are large enough, or announced clearly enough, for creators and businesses to notice directly.
How to tell an update happened, versus normal variance
A genuine algorithm update tends to show up as a broad, sustained shift affecting many accounts at once, often discussed widely across an industry within days. A single account's bad week is far more often normal variance or a change specific to that account's own content, not evidence of a platform-wide update.
The right response is usually smaller than it feels like it should be
Most updates shift emphasis rather than invalidate an entire strategy. A measured adjustment, like trying more of a format that seems to be favored now, tends to be more effective than abandoning an approach that was working before the update. Reacting to every reported update with a full strategy overhaul produces more churn than improvement.