Did the migration lose pages?

The single most expensive migration failure is the quiet one: a page that existed on the old site and simply isn’t there on the new one, discovered weeks later when someone notices traffic to it went to zero. This answers exactly that question, fast, before the full gate runs.

1. Run the source-versus-target comparison against both public hosts before cutover, while the old site is still reachable to compare against.
2. Read every row the report marks missing, moved, or materially changed — title, canonical, body length and structured-data types all count as “changed,” not just a vanished URL.
3. [manual] Have the migration owner classify each difference: intended, needs a redirect, or a genuine defect. The tool can see that something changed; it has no way to know whether that was the plan.
4. Fix the confirmed defects on the target, rerun the comparison, and keep the final clean report with the launch record — it’s the evidence that “nothing” is the accurate count, not an assumption.

This is one question, not the whole gate

A clean run here answers “did we lose pages,” and stops there. It says nothing about redirects, forms, or analytics continuity — those are real, separate failure modes with their own checks, and can-this-go-live is where they’re all run together before an actual cutover.