Article

Jun 9, 2026

The Website Redesign SEO Checklist: Relaunch Without Torching Your Rankings

Most redesign checklists are 50-item dumps. This one is ordered by failure cost, sourced from Google's own migration docs, and built for operators who actually have to ship

Single thin line of light bisecting darkness, broken at center with orange glow

Redesigns lose traffic for one reason most of the time: URLs changed and the redirects were wrong, incomplete, or pulled down too early. The design has almost nothing to do with it. If you keep the URL structure intact, map every change to a single 301, and avoid relaunching during a Google core update, you can ship a full visual rebuild on a Tuesday and watch rankings hold flat by Friday.

This is the website redesign SEO checklist we actually run for clients at Entropy, ordered by what costs you the most when it goes wrong.

TL;DR

  • Rankings drop after relaunch because URLs changed without clean 301 redirects, not because the design changed.

  • Crawl the old site and build a URL-to-URL redirect map before a single new page goes live.

  • Google's own docs require 301 redirects stay live for at least 180 days, often longer.

  • Never relaunch during a core update rollout — the May 2026 update ran 13 days and made diagnosis nearly impossible.

  • Keep URLs, H1s, and internal link patterns identical wherever the redesign doesn't require changing them.

1. Where redesigns actually lose traffic (it's URLs, not design)

When a client tells me their rankings tanked after a relaunch, I open three tabs: the old sitemap, the new sitemap, and Search Console's coverage report. Nine times out of ten, the story is in those tabs within four minutes. URLs moved. Redirects went to the homepage instead of the equivalent new page. Or worse, redirects went nowhere and the old URLs return a 404.

Google is explicit on this. Their 301 redirects documentation states that 301 and 308 permanent redirects do not cause PageRank loss, and recommends pointing each old URL directly at its new destination rather than chaining hops. That word directly is doing real work. A chain of three redirects — old URL → temporary URL → category page → final page — leaks signal and slows crawl, even when each hop is technically a 301.

The design system, the new color palette, the headless CMS migration — none of it matters to Googlebot if the URL contract holds. Keep the contract, ship the new design.

2. Pre-launch: full crawl, URL map, traffic benchmark

Three artifacts have to exist before staging is even spun up.

The full crawl. Run Screaming Frog or Sitebulb against the live site and export every indexable URL, its status code, its canonical, and its top three internal linking pages. For a 400-page services site this takes about 45 minutes. For an e-commerce catalog with 12,000 SKUs, budget a half day and a paid license.

The URL map. A spreadsheet with two columns: old_url and new_url. Every row from the crawl gets a destination. If a page is being deleted, the row gets the closest topical parent — the service category, not the homepage. Homepage-as-catchall is how you lose 30% of long-tail traffic in a weekend.

The traffic benchmark. Pull the last 90 days of Search Console data: clicks, impressions, average position, and the top 200 landing pages. Export it. Save it twice. This is the only baseline you'll have when someone asks in week three whether the redesign "hurt SEO," and you need to answer with a number.


Sequence from pre-launch crawl through 180-day redirect hold

The relaunch sequence — each stage gates the next.

3. The redirect rules Google itself publishes

Most checklists invent redirect best practices. Google already published them. Two documents are load-bearing:

The site move with URL changes guide says to keep redirects live for at least 180 days, and longer if Search traffic is still hitting the old URLs. That's the floor, not the ceiling. We've kept redirects live for 14 months on sites with deep backlink profiles because the logs showed Googlebot still requesting old paths monthly.

The 301 redirects guide confirms two operational facts: permanent redirects pass full signal, and chained redirects should be collapsed to single hops. If your migration has a chain like /old-blog/post → /blog/post → /resources/post, fix the source so the first hop lands on the final URL.

A few rules we add on top, from running these projects:

  • Test the redirect map on staging before launch. Take 50 random old URLs, hit them on the staging host, confirm each returns a 301 to the correct new URL.

  • Use server-level redirects (nginx, Apache, Cloudflare Workers) not meta refreshes or JavaScript redirects.

  • Keep the redirect file in version control. The intern who "cleans up" the redirect config in month four is a real risk.

4. What to keep identical when you can: URLs, headings, internal links

The redesign brief usually says "modernize the site." The SEO read of that brief is: change the visual layer, freeze the structural layer.

  • URLs. If /services/agentic-ai-consulting ranks today, the new site should serve that exact URL. Not /what-we-do/agentic-ai-consulting. Not /services/agentic-ai. The same string.

  • H1s. If the current H1 is "Agentic AI Consulting for Mid-Market Operators," don't rewrite it to "AI That Works" because the new design has a shorter hero. Keep the H1, restyle it.

  • Internal links. The patterns that connect blog posts to service pages are part of why those service pages rank. Audit your top 20 internal link sources before launch and confirm each new template preserves them. We wrote more on this in the conversion-focused redesign process.

  • Page titles and meta descriptions. Carry them across verbatim for the top 50 traffic pages. Rewrite the long tail later, once you have post-launch data.

There's a related decision that affects this — whether to consolidate service pages into one or keep them separate — which we covered in one page vs separate service pages. Resolve that question before the URL map, not after.

5. Launch week: what to watch in Search Console

The first seven days post-launch are the diagnostic window. Three reports matter:

The Coverage report will show new "Not found (404)" entries within 48–72 hours if redirects are missing. Check it daily. Any old URL appearing there as a 404 is a row missing from your redirect map. Fix the row, request validation.

The Page indexing report filtered by submitted sitemap tells you whether the new URLs are being discovered. If you submitted a new sitemap and seven days later only 40% of pages are indexed, the issue is usually either a stray noindex tag from staging or a robots.txt rule that didn't get updated.

The Performance report with the date comparison set to "last 7 days vs previous 7 days" surfaces the URLs losing impressions fastest. Sort by impressions delta. The top 10 losers are your priority queue — open each, walk the redirect, check the rendered HTML, look for missing canonical tags.

Do not panic-edit content in launch week. The signal-to-noise ratio is terrible. Triage redirects and indexation first; touch copy in week three.

6. Timing: never ship during a core update rollout

This is the rule competing checklists omit, and it's the one that ruins post-launch diagnosis.

Google's May 2026 core update rolled out from May 21 to June 2 — 13 days of search volatility, more pronounced than the March update before it. If you relaunched on May 25, you have no way to tell whether a traffic drop in week one came from your URL map, a missing canonical, or the algorithm reweighting topical relevance across your category. The variables are entangled.

The operating rule we give clients: before scheduling a relaunch date, check the Google Search Status Dashboard for any in-progress ranking system update. If one is rolling out, push the launch by two weeks past the announced completion date. If a core update gets announced after your launch is scheduled, the same rule applies — slip the date.

In practice this means treating the rollout calendar as a constraint on your project plan, not a footnote. The cost of a two-week slip is much smaller than the cost of debugging an attribution problem for a quarter.

7. The 180-day rule and when you can finally relax

Google's site move documentation sets 180 days as the minimum redirect lifetime — about six months. That's the regulatory minimum. The operational version is: keep redirects live until your server logs show fewer than 10 monthly requests to old URLs from Googlebot, and zero from real users via referral links.

For most sites this happens between months 8 and 14. Sites with strong backlink profiles in academic or news categories sometimes never reach the threshold and run redirects indefinitely. That's fine. Redirects are cheap. Lost authority is not.

A few signals that tell you the migration is done:

  • Search Console impressions on the new URLs match or exceed the old baseline for 30 consecutive days.

  • The Coverage report shows zero new 404s related to old paths for two weeks.

  • Server logs show old URL requests dropping to near-zero for both bots and users.

When all three line up, you can archive the redirect spreadsheet — but leave the rules in production. Pulling them at month seven because "the migration is over" is how sites lose 20% of traffic in October when a backlink from 2021 finally gets crawled again.

FAQ

How long does a website redesign typically take if you do SEO properly?

In our client work, a content-stable redesign with a clean URL map takes 8–12 weeks end to end. About two of those weeks are dedicated to the crawl, redirect mapping, and staging redirect tests before any new design ships. Skipping that pre-work shaves a week off the timeline and tends to cost 15–30% of organic traffic.

Will I lose rankings if I redesign my website without changing URLs?

If URLs, H1s, and internal link patterns stay identical, ranking impact is typically minimal — sometimes positive once Core Web Vitals improve. The variables that move rankings during a redesign are URL changes, canonical errors, accidental noindex tags carried over from staging, and content rewrites. Pure visual changes rarely affect organic performance.

Do I really need 301 redirects for every old URL?

Yes, for every URL that was indexed or had inbound links. Google's 301 documentation confirms permanent redirects pass full ranking signal, but only if each old URL points directly to its equivalent new URL. Skipping low-traffic pages is a false economy — many earn their links from a single high-authority source you don't want to lose.

How do I know if a core update is happening before I relaunch?

Check the Google Search Status Dashboard before locking your launch date and again 48 hours before launch. Google announces core updates on the dashboard and on X via @googlesearchc. If a rollout is in progress or scheduled, push your launch two weeks past the announced completion date to keep diagnosis clean.

What's the single biggest cause of a relaunch website rankings drop?

Incomplete or incorrect 301 redirect mapping — specifically, old URLs that either return 404s or redirect to the homepage instead of the topically equivalent new page. This single failure mode accounts for the majority of the post-launch traffic drops we've diagnosed. The fix is the redirect map built before launch, not after the drop shows up in Search Console.

Operator close

This week: run a full crawl of your current site and export it to a spreadsheet. That single artifact is the foundation every other step depends on, and most teams never produce it until something has already broken.

If you want a second set of eyes on the redirect map before you ship, send it over.

© All right reserved

© All right reserved