Most relaunches are scoped as design projects with a technical footnote. The footnote is the project. Account for every URL before anything ships and your rankings survive the trip. Skip it and you find out six weeks later, in the traffic graph.
Every migration that ends badly starts the same way. The project was called a redesign. The kickoff deck was about brand. Somewhere near the end, one bullet mentioned that the site would also be changing platforms, as if that were a detail.
It is not a detail. Your standing with search engines is attached to your URLs. Every ranking, every backlink, every bookmark, every link in a customer's onboarding doc points at an address. Move the addresses without a forwarding plan and you are asking Google to re-evaluate your business from scratch, at the exact moment you have the most to lose.
Before design starts, crawl everything. Not the pages in the sitemap. Everything: the landing pages marketing forgot, the PDFs sales still emails, the blog posts from years back that quietly own a keyword. The count is usually a surprise, and the surprise is usually large.
SEEN IN THE FIELD [REPRESENTATIVE]: 2,300 URLS CRAWLED / 214 CARRYING TRAFFIC / 41 HOLDING RANKINGS THAT PAY
Not every URL deserves to survive the move. Some should be retired, and retirement is fine. But retirement is a decision, made per URL, with the traffic data on the table. Orphaning is what happens when nobody decides.
The redirect map is not a spreadsheet chore. It is the part of the project where the revenue lives.
A flat list of URLs is not an inventory. Each one needs a value: what traffic it carries, what links point at it, what rankings it holds, what pipeline those rankings feed. The grading is what turns the map from a compliance exercise into a set of priorities. The forty URLs that pay get moved like freight. The four hundred that never earned a visit get an honest burial.
The map gets tested on staging with the same rigor as the design review. Every legacy URL, requested and followed, until the chain resolves where it should. Once is not enough. Maps drift as content changes during a build, so it gets tested again the week of launch.
Then watch for ninety days. Rankings wobble after a move even when everything was done right. The difference between a wobble and a collapse is whether anyone is looking, and whether the person looking can tell the difference.
None of this is glamorous. That is the point. A migration done well is a non-event, and a non-event is the best thing a launch can be.
One email when something worth reading exists. No cadence, no drip, unsubscribe stays one click.