Most migration traffic losses are not caused by one dramatic error. They come from assumptions nobody checked before launch.
Quick answer
The costly mistakes are usually these: scoping the move from what you think the site contains instead of what it does contain, trusting sitemaps and staging code over the live site, changing too many things at once, redirecting carelessly, and not measuring afterwards. The fix is boring and reliable: crawl the live site, record a baseline, keep URLs and on-page signals the same wherever you can, and check the result against the baseline after launch.
What I found moving my own site
I recently moved peacocksearch.co.uk from WordPress to a Next.js build on Vercel. It is a small site, but a careful pre-migration crawl still found things I would have got wrong by assuming.
- The site was more than twice the planned size. The plan assumed about 40 URLs. The crawl found 101 indexable URLs, including 53 Search Notes, 11 industry pages and six tools.
- Two sitemaps were live. A custom sitemap listed 99 URLs, and the Yoast sitemap listed 101, including two duplicate URLs that canonicalised elsewhere.
- The theme code did not match the live site. The theme files disabled the Yoast sitemap and redirected one of the duplicate URLs. The live server did neither. I treated the live site, not the code, as the source of truth.
- Settings lived in code, not content. The analytics ID, security headers, form handling and a redirect-checking tool were all in theme files. None of it would appear in a content export, and all of it needed rebuilding.
- Structured data was already there. An earlier audit had said there was none. The crawl showed schema on 93 of 101 pages, so the rebuild had to reproduce it, not start again.
- Old bugs came across too. The old pages emitted two canonical tags, and there were no Open Graph tags at all.
None of these was dramatic. Each one could have quietly cost something if I had worked from a checklist instead of from the live site.
The classic mistakes
Not recording a baseline
Before changing anything, I saved a crawl of every live URL: status, title, meta description, H1, canonical and schema types. After launch you compare against it. Without it, "did anything get worse?" becomes an argument from memory.
Changing everything at once
A migration is risky enough on its own. If you also rewrite titles, restructure URLs and change the copy, you cannot tell which change moved rankings. My rule was: keep titles, H1s and body content identical and URLs unchanged wherever possible, and only add things that cannot hurt, such as correct canonical and social tags.
Redirecting everything to the homepage
Each old URL should go to the closest equivalent page, one to one. Google can treat redirects to an irrelevant page, such as the homepage, as soft 404s, and the original page's relevance is lost. See how to build a redirect map.
Leaving staging rules in place
A noindex tag, a password, or a robots.txt block that was right on the staging site can survive into production and remove the whole site from Google. Check these on launch day, not next week. The pre-launch checklist covers it.
Forgetting internal links and sitemaps
Redirects are a safety net, not a plan. Update internal links to point at the final URLs, and make sure the new sitemap lists only the pages that return 200 and are indexable.
Not watching what happens next
A migration is not finished at launch. Watch Search Console for indexation changes and crawl errors, re-crawl the old URL list to catch redirect problems, and compare traffic against the baseline. What to monitor after a website migration lists what to check and when.
What I will not claim
The site is young and has little traffic, so it is too early to say what the migration has done for rankings. The value of the process is the same regardless: you find the surprises before Google does.
When to get support
If you are planning a migration, the cheapest time to get an SEO review is before the URLs change. My SEO migration support covers the pre-launch audit, redirect planning and post-launch monitoring.
FAQ
How much traffic loss is normal after a migration?
Some short-term movement is common. A drop that persists for weeks, or affects your important pages, usually points to a problem with redirects, indexation or content changes, not to normal volatility.
Should I change URLs during a migration?
Only if you have to. Every URL change adds risk. If you do change them, every old URL needs a redirect to its closest new equivalent.
How long should I keep redirects?
At least a year, and longer for URLs that have backlinks or steady traffic.