SEO Migration Checklist for Website Redesigns

A redesign is finished when the new site looks right. A migration is finished about three months later, when search traffic has recovered and you know which pages did not.

Most of the redesigns that lose traffic did not lose it through bad design. They lost it because URLs changed without redirects, because pages that earned traffic were cut as "old content," or because nobody looked at Search Console until the quarter ended.

This is the checklist we run. It is deliberately tedious. The tedium is the point.

Phase 1 — before the redesign starts

This is the phase that gets skipped, and skipping it is what makes recovery impossible later. You cannot restore what you never recorded.

1. Export every current URL. Crawl the site with Screaming Frog or similar, and cross-check against your CMS, your sitemap, and server logs. Include images, PDFs, and anything else that gets linked. Google's guidance is to build the list from sitemaps and analytics, then verify against the CMS and logs, because each source misses something the others catch.

2. Identify pages receiving organic traffic. Search Console → Performance → Pages, over the last 12 months. Twelve, not three — seasonal pages will not appear in a shorter window.

3. Record current rankings and queries. Export query-level data per page. This is your baseline. Without it, "traffic is down" is an argument rather than a diagnosis.

4. Find your backlinks. Any link tool, plus Search Console → Links. Pages with external links are the ones where a broken redirect costs most — you lose the link value, not just the page.

5. Record metadata. Titles, meta descriptions, H1s, canonical tags, and structured data for every page that matters. Rewriting titles wholesale during a redesign is a common, avoidable cause of traffic loss.

6. Note what is actually indexed. Search Console → Pages. Indexed pages, excluded pages, and why. Pages Google has chosen not to index do not need redirects.

7. Identify high-converting pages. From analytics and the CRM. These outrank traffic in importance. A page with 200 visits and 12 inquiries matters more than one with 4,000 visits and none.

Save all seven exports in one folder, dated. You will need them.

Phase 2 — during the redesign

8. Preserve the content that works. Decide from the data in phase 1, not from how the content looks in the new design. If a page ranks and converts, its content survives the redesign — the presentation can change.

9. Plan the new URL structure. The best migration changes no URLs at all. If you must change them, make the change once, deliberately, and keep the structure consistent.

10. Map old URLs to new URLs, one by one. Every old URL gets exactly one destination. This is the core artifact of the whole migration.

Old URL	New URL	Type	Organic clicks (12mo)	Backlinks	Notes
/services/boiler-repair	/services/boiler-repair	Unchanged	1,240	8	No action
/plumbing-sheffield	/services/plumbing/sheffield	301	380	2	Structure change
/blog/old-post	/blog/new-consolidated-post	301	95	1	Merged
/team/former-staff	/about	301	4	0	Nearest relevant
/tag/misc	—	410	0	0	Deliberately removed

Two rules. Redirect to the closest relevant page, not the homepage — bulk redirects to the homepage are treated as soft 404s and pass nothing on. And for pages that genuinely should not exist any more with no relevant destination, a 410 is honest and clearer than redirecting someone somewhere useless.

11. Maintain internal linking. Update internal links to point at the new URLs directly rather than relying on redirects. Links that hop through a redirect still work, but they add latency and they make future auditing harder.

12. Preserve heading intent. Keep H1s and section headings semantically close to what ranked. A page that ranked for "emergency boiler repair" and now leads with "Heating Solutions" has changed what it is about.

13. Carry structured data across. Anything you had — LocalBusiness, FAQPage, Article, BreadcrumbList — needs rebuilding on the new templates. It disappears silently otherwise.

Phase 3 — before launch

Run this on staging, with the staging site blocked from indexing.

14. Implement the redirects. Server-side permanent redirects. Google recommends HTTP permanent redirects "such as 301 and 308" where technically possible. Avoid chains — Googlebot will follow up to ten hops, but each one costs latency for real users, and chains longer than a few hops are a maintenance problem.

15. Test every form. Every form, every destination, every notification. Forms break in redesigns constantly, and a broken form is invisible until someone notices the inbox has gone quiet.

16. Check canonical tags. Each page should be self-canonical unless there is a specific reason otherwise. A staging canonical shipped to production is a classic and severe launch bug.

17. Check robots directives. The one to verify twice: noindex and Disallow: / from staging must not reach production. This is the single most damaging migration mistake there is, and it is entirely avoidable by checking.

18. Generate and validate the sitemap. New URLs only, all returning 200, no redirected or blocked URLs in it.

19. Verify analytics. Tracking code present on every template, conversion events configured and firing, consent handling correct. See GA4 events every lead-generation website needs.

20. Prepare Search Console. Verify all property variants — www and non-www, HTTP and HTTPS. For a domain change, be ready to use the Change of Address tool.

Phase 4 — after launch

21. Test the redirects for real. Crawl your full list of old URLs against production. Every one should resolve to its mapped destination in a single hop with a 200 at the end. Do this within hours of launch, not days.

22. Submit the new sitemap. And keep the old one submitted for a while — watching indexed pages migrate from the old sitemap to the new one is a genuinely useful progress signal.

23. Monitor 404s. Search Console → Pages → Not found, plus your server logs. Anything appearing here is a URL you missed. Expect to find some; the list should shrink each week.

24. Monitor indexing. Watch the new URLs get indexed and the old ones drop out. Google notes that for a medium-sized site "it can take a few weeks or more" for new URLs to replace old ones in results.

25. Monitor traffic and conversions. Both. A migration that holds traffic and halves inquiries is a failed migration, and it is entirely possible if a CTA or a form moved.

26. Compare old and new landing pages. Page by page against the phase 1 baseline. Aggregate numbers hide the interesting cases — one page collapsing while another gains can net out to "flat."

Launch day QA

Check	Pass condition
robots.txt	Production version, not staging's
Meta robots	No stray noindex on live pages
Canonical tags	Self-referencing, production domain
Redirects	All mapped URLs, one hop, 200 at the end
Homepage redirect	Old homepage variants resolve correctly
HTTPS	Forced; no mixed content
Sitemap	Live, valid, new URLs only
Search Console	All variants verified, sitemap submitted
Analytics	Firing on every template
Conversion events	Verified with a real test submission
Forms	All submit and deliver
Phone numbers	Correct and tappable
Structured data	Validates on key templates
404 page	Exists, returns 404, links somewhere useful
Page speed	Checked on top templates
Internal links	Point at new URLs directly
Images	Loading, with alt text preserved

The first 30 days

Day 1: Full QA above. Crawl old URLs. Submit sitemap. Test forms from an external network.

Days 2–3: Check Search Console for crawl errors. Watch indexing begin. Test forms again — genuinely, they break twice.

Week 1: Daily 404 checks. Compare traffic against the same week last year, not last week. Fix any redirect gaps as they surface.

Week 2: Index coverage review. Query-level comparison on your top twenty pages. Expect movement; look for pages that have fallen out entirely rather than shifted.

Week 3: Conversion review — are inquiries arriving at the same rate? Check Core Web Vitals field data as it starts to populate.

Week 4: Full comparison against the phase 1 baseline. Produce a page-level list of what is up, flat, and down. Decide what needs intervention.

Months 2–3: Monthly reviews. Most recovery happens in this window. If a page has not recovered by month three, look at it individually rather than waiting longer.

What usually goes wrong

Staging directives shipped to production. The most damaging and the most preventable.

Bulk redirects to the homepage. Treated as soft 404s. Passes essentially nothing. Map individually.

Content cut because it looked dated. Check the data before deleting anything. Old posts are often the ones earning links.

Titles rewritten wholesale. Changing the design is not a reason to change every title. Change them deliberately, with a reason, and watch what happens.

Tracking broken silently. The site works, the numbers quietly stop being real, and you spend a month debating a traffic drop that is a measurement artifact.

Nobody assigned to monitoring. The project ends at launch, the agency moves on, and the drop is discovered in the quarterly review when the cause is three months cold.

SEO & Search VisibilityWe handle redesign structure, development, SEO migration, tracking, and launch QA as one project, so the search work is not a handover.→

Related reading: Website Redesign Cost in 2026 and WordPress vs Framer in 2026.

Frequently asked questions

How much traffic will we lose?

It depends on how much changes and how well the migration is executed. A redesign that keeps URLs, content, and structure intact often shows little change. A platform move with new URLs and rewritten content usually dips and takes weeks to months to settle. A migration with no redirects can lose most organic traffic and not recover.

Should we keep the same URLs?

Yes, wherever possible. Unchanged URLs are the lowest-risk migration there is. Change them only when the existing structure is genuinely blocking something — and then change them once.

How long do we keep redirects?

At least a year, per Google's guidance, and indefinitely if it costs you nothing. Redirect maps are cheap to keep and expensive to reconstruct.

Can we redirect everything to the homepage?

No. Google treats mass redirects to an irrelevant page as soft 404s, so they pass essentially no value and give the visitor a worse experience. Map to the closest relevant page, or return a 410 where nothing relevant exists.

When is the best time to launch?

Early in the week, in the morning, when the people who can fix things are available. Never on a Friday afternoon, and not immediately before your busiest season — you want a normal period to observe, and the ability to respond within hours.

---

If you are planning a redesign, replatform, or domain change, talk to us before the build starts — the URL mapping and baseline work costs far less to do upfront than to reconstruct after traffic drops.