Is It Out of Scope? — Web development
Is fixing bugs found after sign-off out of scope for a web development project?
It depends on the cause. A defect that traces back to the build itself is typically covered by warranty; a bug introduced by something that changed afterward — a dependency update, the client's own code changes, a third-party integration — is a separate, billable investigation and fix.
The build passed QA, the client tested it, and it was formally accepted. Later, a bug surfaces — a form silently failing, a calculation returning the wrong result, a page erroring out for some users. The reflexive assumption is often that anything called a "bug" is the agency's problem to fix for free, but that's only true if the defect actually originated in the delivered code rather than in something that happened to the environment afterward.
Is it out of scope?
It depends on root cause, which is worth actually investigating before promising anything. A defect that was present in the code at sign-off — missed in testing, a genuine mistake — is typically covered under warranty. A bug caused by a dependency or platform update, a change the client's own team made, or a new integration added after launch is a new issue outside the original build, and diagnosing and fixing it is billable.
Clause typically implicated
Clause typically implicated
Warranty clause→ — Defines what counts as a covered defect versus a new issue, and the window during which covered defects are fixed at no charge — root cause is usually the deciding factor.
Example contract wording
Example contract wording (illustrative, not legal advice)
I'll dig into this before quoting anything. If it traces back to the original build, it's covered under our [X]-day warranty and I'll fix it at no charge. If it turns out to be caused by [a dependency update / a change on your end / the new integration], I'll let you know what I find and provide a separate estimate for the fix.
How MarginFlow flags it
MarginFlow weighs the warranty terms, the time elapsed since sign-off, and anything in the thread suggesting the environment changed since launch — a bug reported the day after sign-off reads very differently than one reported three months later with an intervening "we added X" message. Where the cause is genuinely unclear from the email alone, the classifier lowers its confidence and flags the ambiguity explicitly rather than guessing.