Do I need to test again after each update to my site?

Accessibility is an ongoing requirement, not a snapshot

Yes, in practice it is. WCAG 2.1 level AA, as defined in EN 301 549, is not a seal of approval that is achieved once and then remains valid — the standard describes how a site must function at all times while it is online. A new checkout button, a modified checkout flow, a different form, or a new image can affect accessibility again, regardless of how clean the previous test was.

That does not mean every text change requires a full test. What does apply: changes that affect the interaction, navigation, or structure of a page — a different form field, an updated error message, a new step in the checkout process — typically touch precisely the requirements that WCAG 2.1 AA sets for things like keyboard operation, error handling, and reading order. Anyone who tests only at launch and then stops monitoring the site runs the risk that a later change takes the site out of compliance again, even if there is an accessibility statement claiming otherwise.

Where the ongoing requirement comes from

Directive (EU) 2019/882 refers to EN 301 549 V3.2.1 as the standard to which the service must comply, and describes requirements for the end result — not for a fixed test moment. This makes the standard structurally applicable, not as a snapshot. An automated scan, by the way, never shows this conclusively: it detects some of the errors that can be found by machine, but whether alternative text truly describes an image or a checkout step remains logical with keyboard only is beyond the scope of such a scan — at the first test and at every test thereafter. An overlay placed on top of the site afterwards does not change the underlying page and does not meet the requirement, even after an interim change.