Start with a reproducible failure
Record URL, version and time. Confirm the change inside the built file and correct deployment environment before investigating caching. Compare delivered HTML and relevant headers through authorised access.
Use a small controlled test to separate a symptom from its cause. Record the starting state, affected URL and observed outcome so that another person can repeat the check.
A fictional website example
Fictional example: A course library changes a heading, but the build uses old source data. Purging cannot create text that was never built.
Collect evidence before changing anything
New build, old public response: compare origin and intermediate caches. New profile correct, old profile wrong: inspect local reuse and any service worker. A new query parameter is a separate case.
| Check | Expected outcome |
|---|---|
| Starting condition | New CMS text does not prove the right version was built or delivered. Follow the same marked change from build artifact through origin, shared cache and browser. |
| Implementation | Record URL, version and time. Confirm the change inside the built file and correct deployment environment before investigating caching. Compare delivered HTML and relevant headers through authorised access. |
| Verification | The same marked change appears in the build file and delivered page for the new and returning visits included in the test. |
Fix the cause and repeat the test
Correct the verified layer. Test ordinary loading and returning visits at the normal URL after the next publication. A successful purge does not prove a browser copy disappeared.
Repeat the same test after the change and add a nearby case that already worked. One successful case establishes that task in those conditions; it does not certify the whole website.