Everyone knows migrations need redirects. Programmes still lose thirty per cent of their traffic to them, and it is almost never because someone forgot the redirects entirely.
The map is built too late
Redirect mapping is treated as a launch-week task. By then the new URL structure is fixed, the old inventory has not been fully exported, and whoever builds the map is working under deadline against pages they did not create. Build it during design, when the structure can still change to accommodate what already ranks.
Chains are where equity quietly leaks
Every migration adds a hop to the previous one. After three site moves, an old URL passes through four redirects before landing. Each hop adds latency and each one is a chance for something to break. Rewrite the map from the original URLs each time, not from the current ones.
Redirect to the equivalent page, not the homepage
Bulk-redirecting retired pages to the homepage is the standard shortcut and it wastes almost everything those pages earned. A redirect to an irrelevant destination is treated much like a soft 404. Where no equivalent exists, the honest answer is a 410 and a plan to rebuild.
What to monitor, and for how long
Watch indexed page counts, crawl errors and the top hundred landing pages weekly for a full quarter. Migration damage rarely appears in week one — it shows up around week three, once recrawling has worked through the site, by which point the launch team has usually moved on.