Protect the search equity your new website cannot recreate overnight

Phase 1: Establish the baseline

Capture the state of the current site before staging, DNS, analytics, or URLs change. The baseline is both a migration input and the reference used to diagnose a post-launch movement.

• Crawl every accessible HTML page, PDF, image, canonical, redirect, and error URL.
• Export at least 16 months of Search Console page, query, country, device, and search appearance data when available.
• Export organic landing pages, conversions, engagement, and revenue or pipeline attribution from analytics and the CRM.
• Export backlinks and identify linked URLs that no longer appear in navigation.
• Record current robots.txt, XML sitemaps, canonical tags, structured data, hreflang, and important headers.
• Save representative Core Web Vitals and performance measurements by template.
• Identify branded and non-branded query clusters and the page currently serving each intent.

Phase 2: Build the URL inventory and disposition map

Combine crawl, sitemap, analytics, Search Console, backlink, CMS, server-log, and stakeholder sources. A page absent from the current navigation may still have links, traffic, contractual value, or a useful historical ranking.

For every old URL, record the proposed new URL and one disposition:

Phase 3: Define the new site's search architecture

• Assign one primary search intent to each commercial or editorial page.
• Use lowercase, readable, durable URLs with one consistent trailing-slash policy.
• Keep valuable pages reachable through crawlable HTML links and descriptive anchor text.
• Separate service, industry, work, insight, team, and location content by purpose rather than repeating the same sales page.
• Avoid thin city or industry variants that simply change place names and funnel visitors to the same destination.
• Design visible breadcrumbs and useful parent-child relationships.
• Plan canonical handling for filters, parameters, print views, campaign URLs, and other duplicates.

Phase 4: Make staging safe and testable

Protect staging with authentication whenever possible. A robots.txt block alone does not reliably keep known URLs out of search, and it can prevent crawlers from seeing a noindex directive. Remove every staging-only control at launch.

• Render meaningful page content and internal links in crawlable HTML.
• Confirm one indexable canonical version for protocol, hostname, casing, slash, and default-document variants.
• Generate unique titles, descriptions, headings, canonicals, and social metadata by template.
• Validate structured data against visible page content and use stable entity identifiers.
• Make pagination crawlable; do not rely on infinite scroll as the only discovery route.
• Ensure search, filter, preview, account, and utility URLs have an intentional indexing policy.
• Test error templates for genuine 404 responses and maintenance behavior for 503 responses.

Phase 5: Validate content and on-page signals

• Confirm the new page satisfies the same or a deliberately improved intent before mapping an old URL to it.
• Preserve useful copy, media, downloadable resources, title concepts, headings, and internal links during rewrites.
• Update internal links to final destinations rather than relying on redirects.
• Carry forward image alt text, captions, dimensions, and image URLs when the asset remains valuable.
• Verify author, publication, modification, client, and location claims before marking them up.
• Keep legally or operationally required documents discoverable and current.

Phase 6: Test the redirect map

Test the map as data before it becomes server configuration. Each source should be unique, each destination should return 200, and each redirect should reach its final destination in one hop whenever possible.

• Test HTTP, HTTPS, www, non-www, uppercase, lowercase, and legacy default-file forms.
• Test query-string and file-extension patterns only where a rule is intentional.
• Review high-value URLs manually; do not trust pattern rules without sampling exceptions.
• Keep redirects for at least a year and preferably as long as people or external links still use the old URLs.
• Do not include redirected, blocked, noindex, error, or noncanonical URLs in the new sitemap.

Phase 7: Prepare measurement before DNS changes

• Install and validate GA4 or the selected analytics platform in production-safe test mode.
• Measure successful submissions, not button clicks, as lead events.
• Capture form starts, project-form completions, calls or email actions where relevant, and key resource engagement.
• Connect lead source, landing page, and campaign data to CRM stages such as generated, qualified, working, and won.
• Verify a Search Console domain property and retain useful URL-prefix properties.
• Annotate launch and major follow-up changes in the reporting record.

Launch day: Control the sequence

1. Take a final backup and confirm the rollback owner and threshold.
2. Deploy the production build and remove staging noindex or authentication controls only from the public destination.
3. Activate and sample the redirect map, prioritizing historically valuable URLs.
4. Check robots.txt, canonicals, response codes, sitemaps, structured data, forms, analytics, and critical integrations.
5. Crawl the live site from an empty queue to expose internal 404s, redirect links, orphan pages, and unexpected canonicals.
6. Submit the canonical sitemap in Search Console and inspect representative page types.
7. Update owned external destinations such as business profiles, social profiles, ads, email templates, and partner listings.

After launch: Watch leading indicators first

Rankings can fluctuate while search engines recrawl and reassess a changed site. The first priority is distinguishing expected processing from implementation defects.

• Daily at first: uptime, server errors, 404s, form delivery, integration failures, robots, and critical redirects.
• Several times in week one: sitemap processing, indexing reasons, selected canonicals, crawl requests, and representative URL inspections.
• Weekly through stabilization: clicks, impressions, non-branded query groups, landing pages, conversions, and qualified leads compared with the baseline.
• Monthly: technical health, content cluster performance, backlink reclamation, pipeline contribution, and remediation priorities.

Investigate patterns rather than isolated positions. A drop limited to one template, directory, device, country, or query cluster often points to a specific cause. Record every fix and its deployment date.