How to Prepare an SEO Migration Before a Website Redesign

Most migration losses are decided before design starts

When a redesign loses search traffic, the post-mortem usually finds a missing redirect file or a staging block left in place. Those are real failures, and the website migration SEO checklist covers how to execute a migration so they do not happen. But they are rarely the root cause. The root cause is usually that nobody assessed the search risk when the project was scoped, so nobody budgeted for it, assigned it or built it into the timeline.

This article is about that earlier stage: the planning a business owner or project lead should do before the RFP goes out or the agency is briefed. It is less about tasks and more about decisions. Which changes are worth the risk, what you need to measure before anything moves, what you will protect at all costs, and who is accountable.

If you are also planning what content moves to the new site, the website content migration plan covers the inventory and rewriting side. The two plans should be written together, because content decisions and search decisions are the same decisions seen from different angles.

Classify how much is changing

"Redesign" covers everything from a new colour palette to a new company name on a new platform. Search risk scales with how many fundamentals change at once. Before anything else, write down which of these apply:

• Visual design only. Same platform, same URLs, same content. Low risk, as long as content and internal links are not quietly removed in the process.
• Content and navigation. Pages rewritten, merged or cut, menu restructured. Moderate risk, mostly from lost pages and changed internal links.
• URL structure. Page addresses change, even on the same domain. Every changed URL needs a redirect and a search engine has to reprocess it.
• Platform or CMS. New system, often with different templates, rendering, metadata handling and default behaviours.
• Domain. A new domain name, often following a rebrand or merger. The highest-risk change, because search engines have to transfer their understanding of the whole site.
• Language or regional structure. Adding French, splitting by province, merging regional sites. Significant risk for bilingual organizations in New Brunswick, where language versions need to be linked correctly from day one.

Each layer adds risk, and they compound. A rebrand that moves to a new domain, a new platform and a new URL structure while cutting half the content is several migrations at once. Sometimes that is unavoidable. Often it is not, and the most useful planning decision is to separate them: keep URLs stable during the platform change, for example, and restructure later once the new site has settled. That decision belongs at the scoping stage, because it changes the budget and the timeline.

Capture benchmarks before anything changes

You can only show that a migration went well, or diagnose what went wrong, if you know where you started. And once the old site is gone, so is your ability to measure it. Capture benchmarks as soon as the project is approved, and refresh them shortly before launch.

The benchmark set for most established organizations:

• Organic sessions and conversions by landing page, over at least twelve months so seasonal patterns are visible. For a ski hill, a tourism operator or a tax practice, a shorter window can be badly misleading.
• Search Console performance: clicks, impressions and average position by page and by query, exported, not just viewed.
• Your priority queries: the commercial searches that matter most, with current positions noted.
• Indexed page count and any existing indexing or coverage problems, so you do not blame the migration for issues that predate it.
• Pages with external links, from whatever backlink source you use.
• Local listings and Business Profile links, which often point at specific pages that will move.
• Page experience measures such as Core Web Vitals, to compare against the new build.
• Tracking configuration: what is measured and how, so the new site measures the same things. A conversion tracking plan written before launch prevents the common situation where traffic seems to fall because tracking broke.

Store the exports somewhere the organization owns, not only in a supplier's account.

Identify the pages that carry the value

On most established sites, a relatively short list of pages carries most of the organic visits, conversions and external links. Find that list early and treat it differently.

Rank every URL by a combination of organic traffic, conversions, external links and commercial priority. Then set a default rule for the top tier: these pages keep their URLs and their substance unless there is a strong, written reason to change them. Design can change. Wording can be improved. But the core content and the address that search engines and other sites know should not be thrown away casually.

This rule creates useful friction. When a designer or stakeholder proposes merging two high-value service pages, or renaming the URL of a page that holds most of the site's links, the decision is made deliberately, with the data in front of everyone, rather than discovered after launch.

For the long tail of low-value pages, the planning question is simpler: keep, merge or retire, with a redirect target for anything that has links or search history.

Assess the specific risks

With the scope classified and the valuable pages identified, write a short risk register. It does not need to be elaborate. For each risk, note how likely it is, what it would cost, and what will prevent it. The usual entries:

• Lost pages. Valuable content cut without a redirect. Prevention: a URL map signed off before build, covering every URL with traffic, links or conversions.
• Changed content on high-value pages. A rewrite that removes the substance that made a page rank. Prevention: review rewrites of top-tier pages against what they currently rank for.
• Platform defaults. A new CMS that generates different URLs, drops metadata, adds duplicate versions of pages or renders key content in ways search engines struggle with. Prevention: a technical review of the platform choice before contracts are signed.
• Staging leaks or launch blocks. A development site indexed, or a launch with search engines still blocked. Prevention: authentication on staging and an explicit launch check.
• Broken tracking. Analytics or conversion events missing at launch. Prevention: tracking specified in scope and tested before launch.
• Language versions misconnected. French and English pages not linked as alternates, or one language left behind. Prevention: bilingual structure specified in the URL map.
• Bad timing. Launching into your peak season, when any temporary fluctuation costs the most. Prevention: choose the window now.

Timing deserves its own line for Atlantic Canadian organizations. A tourism operator launching in June, a ski resort launching in November or an accounting firm launching in March is accepting avoidable risk. Search engines need time to process a migration, and some movement in rankings during that period is normal. Plan to absorb it in a quiet month.

Decisions to make before the project starts

Several decisions shape every later step. Make them in the planning phase and write them down:

1. URL policy. Will URLs stay the same where possible? If the structure changes, what is the new pattern, and is it simple, readable and stable for years?
2. Redirect policy. Every old URL with value redirects permanently to its closest equivalent, one hop, no chains, no blanket redirects to the home page. Who builds the map, who reviews it, and when is it due?
3. Content scope. Which pages are kept, merged, rewritten or retired, decided with search data available rather than by opinion in a workshop.
4. Ownership. One named person accountable for search through the migration, with authority to hold the launch if checks fail. Usually not the designer, and not only the developer.
5. Acceptance checks. What must be true before launch and in the weeks after: redirects tested against the map, metadata carried over, sitemap updated, tracking firing, Search Console verified for the new setup.
6. Post-launch monitoring period. How long the supplier stays engaged after launch, what they monitor, and how fast issues are fixed. The guide to the weeks after launching a new website covers what that period should include.

Put these into the RFP or project scope. A proposal that does not price redirects, a URL map, tracking and a post-launch monitoring period is not cheaper. It has moved that cost onto you, to be paid later in lost inquiries. The website redesign planning guide shows where this fits in the wider project.

A worked example

Consider an illustrative regional insurance brokerage with offices in Fredericton, Saint John and Charlottetown. It plans a rebrand following an acquisition: new name, new domain, new platform and a new bilingual site replacing an English-only one.

The planning review classifies this as the highest risk tier: domain, platform, URL structure and language structure all changing at once. It recommends separating what can be separated. The platform and URL structure change with the redesign; the new domain launches with it because the rebrand requires it; but the French version is planned as a second phase, a few months after launch, so the English migration can stabilize first and the French content can be written properly rather than rushed.

Benchmarks are exported immediately, covering two full years to capture renewal-season patterns. About two dozen pages are identified as the top tier, including the commercial insurance service pages and the office location pages linked from each Business Profile. Their substance is protected, and each gets an individual post-launch check. Launch is scheduled outside the brokerage's busiest renewal period. The URL map, redirect testing, tracking and a post-launch monitoring period are written into the RFP so every bidder prices them.

None of this guarantees a smooth result. It makes a smooth result much more likely, and it means that if something does slip, there is a baseline to diagnose it against.

Migration planning is a standard part of how we approach rebuilds on our SEO and AI visibility work: assessed during discovery, priced in the scope and owned through the stabilization period after launch.