Erika Braeger of Tenspeed has watched enough replatforms go sideways to have a one-line rule, and it is not about redirects. On the Unscripted SEO podcast she put it plainly.
The Question To Ask Before Any Migration
“Never rush a website migration.”
Erika Braeger, Tenspeed
What makes that more than a platitude is where she starts the engagement:
“when I’m working on a website migration with a client, one of the first questions I ask is like, what are your goals?”
Erika Braeger
Not the CMS. Not the URL structure. The goal. Because a surprising number of migrations exist to satisfy a preference rather than a business need, and those are the ones that get compressed into a bad timeline.
Mapping Business Goals To Technical Scope
“What are the business goals behind this migration to make sure that they’re not moving from WordPress to Webflow or vice versa because somebody likes it over the other.”
Erika Braeger
Write the goal down in one sentence, then derive scope from it. “Reduce time-to-publish for the content team” produces a very different project from “support a new product taxonomy”, which produces a different one again from “the marketing director prefers the editor”.
If the sentence cannot be written, the migration is not ready to be scheduled. That single artefact kills more bad migrations than any audit.
The Redirect Inventory Nobody Builds In Time
The redirect map is always started too late, because it is the least interesting deliverable and the one that requires the most complete picture of the old site.
- Crawl the live site and export every indexable URL.
- Pull Search Console for every URL that earned an impression in the last sixteen months — the crawl will miss orphans that still rank.
- Pull server logs for anything a bot still requests.
- Pull the backlink profile so you do not drop an externally linked URL that gets no organic traffic.
- Union all four lists. That is the real inventory.
- Map one-to-one where possible. Flag every URL with no obvious target — those are decisions, not redirects, and they need a human.
Braeger’s other rule follows naturally from this: leaving off entire sections is a failure mode as common as forgetting redirects, and both come from starting the inventory late.
Staging Checks That Catch The Expensive Mistakes
The short list that catches most of the costly errors: robots.txt not carried over from staging, noindex left on templates, canonicals pointing at the staging host, internal links hard-coded to the old domain, structured data silently dropped by the new templates, and pagination or faceted parameters generating a fresh crawl trap.
Run a full crawl of staging with the same configuration you use on production and diff the two. The diff is the deliverable — see technical SEO audit and crawl budget optimisation for the crawl setup, and sitemap optimisation for what to submit on day one.
The Thirty-Day Post-Launch Watch List
Daily for the first week, then weekly: indexed page count against the pre-launch baseline, crawl errors by type, redirect chains that appeared after launch, Core Web Vitals by template, and organic revenue by landing page group.
The last one is the one that gets skipped and the only one an executive cares about. Set the baseline before you launch, or you will spend the recovery arguing about what normal looked like.
Braeger’s wider view on working in the AI era is in her episode recap.