A 10-step, production-tested framework for auditing, migrating, and enhancing a site’s SEO and AEO through a redesign or platform move — without losing the rankings that already work.

Every redesign is a controlled demolition. New templates replace old ones, URLs change shape, copy gets rewritten, and entire sections of the information architecture move. Search engines don’t see a rebrand — they see a site that changed underneath the URLs they already trust. If nothing is done deliberately, the rankings a client has spent years earning are the first casualty.

This is the framework we run on every migration and redesign: audit first, preserve everything measurable, enhance what can be enhanced, and verify at every checkpoint that nothing was lost in translation. It follows the same order every time, because the order is what protects the outcome.

A note on sourcing: claims in this piece are graded [VERIFIED] for documented, stable search-engine behavior, [VENDOR CLAIM] for tool or platform-specific claims worth confirming yourself, and [DISPUTED] for areas where practitioner experience varies.

1. Run the Full SEO Audit — Before Anyone Touches Design

Nothing goes to design until the SEO blueprint is approved. This isn’t a formality; it’s sequencing. Once a designer builds wireframes around a new information architecture, every SEO recommendation after that point is a negotiation instead of a requirement.

A pre-design audit answers three questions: what is currently ranking, why is it ranking, and what happens to that “why” if the page, its URL, or its copy changes. That includes:

Which pages drive organic sessions, conversions, and backlinks — not just which pages rank
Which ranking pages have thin or duplicate content that the redesign should not simply replicate
Which technical issues (index bloat, orphaned pages, crawl traps) should be fixed during the rebuild rather than carried over
Which existing schema, FAQ content, and structured data are already earning rich results or AI citations

If the SEO blueprint isn’t signed off before wireframes start, you’re not protecting SEO — you’re doing damage control after the fact.

2. Get Access, Then Pull the Historical Data

Request admin or read access to Google Search Console and Google Analytics on day one — access delays are the most common reason migration timelines slip. Once in, pull at least 16 months of history (to cover year-over-year seasonality) for:

Top-performing pages by clicks, impressions, and average position in GSC
Indexed vs. submitted URL counts, and any manual actions or security issues
Landing page conversions and goal completions in Analytics, not just traffic
Backlink-driving pages from Ahrefs/Semrush, cross-referenced against GSC’s top pages

This baseline is what every post-launch comparison gets measured against. [VERIFIED]

3. Crawl the Existing Site — Screaming Frog, Ahrefs, Semrush

A full crawl finds what analytics can’t: every URL that exists, whether it’s indexable, its status code, its canonical target, its internal link count, and its current metadata — including pages with no traffic today that still hold backlinks or rank for long-tail queries GSC under-reports.

Screaming Frog: full crawl for URL inventory, status codes, canonical tags, and internal linking structure
Ahrefs / Semrush: backlink profiles per URL, keyword rankings, and organic traffic estimates to cross-check GSC
Cross-reference all three sources — a URL that appears in only one of them is exactly the kind of page a redirect map misses

A redirect map built only from a sitemap will miss the pages that don’t appear in navigation but still rank. This is the single most common cause of post-migration traffic loss.

4. Run the Business Analysis

Rankings alone don’t tell you which pages matter to the business. A page ranking #3 for a low-intent query may be worth less than a page ranking #9 for a query that converts. Understand the client’s services, margins, and priority markets before deciding what the new site architecture should protect, consolidate, or cut.

Which product or service lines are being deprioritized, expanded, or merged in the rebuild
Which markets or languages need dedicated URL paths versus shared content
Which competitor gaps the redesign is meant to close — this shapes new page and content decisions, not just preserved ones

5. Extract Every Existing Page — Copy, Rankings, and URL Structure

Before a single old page is touched, extract it completely: the live copy, the current URL, its ranking keywords, its meta tags, its internal and external links, and its schema. Nothing should be reconstructed from memory or assumption later — it should be pulled from a saved inventory.

Full URL inventory (crawled + GSC-indexed + sitemap-submitted, reconciled into one list)
Page-level copy exports, kept as the source of truth for the rewrite step
Ranking keyword sets per URL, pulled from GSC and rank tracker data
Existing schema markup and structured data per template type

6. Propose the New URL Structure and Lead Keyword/Entity Research

URL structure changes are the highest-risk single decision in a migration. Every folder renamed, every slug shortened, and every category flattened has to map to something — and that mapping is what the redirect plan in Step 8 depends on.

Design the new URL taxonomy around search intent and entities, not just the new site’s navigation labels
Run keyword and entity research before finalizing slugs — not after design has already chosen them
Keep URL changes to what’s necessary; a rebrand does not require renaming every URL on the site

Unnecessary URL changes are the top preventable cause of ranking volatility after a migration. [VERIFIED]

7. Extract Old Copy, Run It Through AI, Propose the Final Copy

AI accelerates the rewrite, but the brief matters more than the model. Feed the AI the extracted copy from Step 5, the ranking keyword set from Step 5/6, and the house writing guidelines — then treat the output as a first draft that gets checked against what the old page was already earning credit for.

Preserve the exact phrases and entities the old copy was ranking for, even while improving structure and readability
Rebuild any FAQ or Q&A content already earning AI citations — don’t let a rewrite silently drop the section a page is being cited for
Score every AI-drafted page against the SEO and copywriting guidelines before it goes to the client, not after

8. Build the One-to-One 301 Redirect Map — Then Confirm at Launch

Every old URL that carries any ranking signal, backlink, or indexed status gets a 301 to its single closest equivalent on the new site. One-to-one, not one-to-home-page. A redirect map that dumps orphaned URLs onto the homepage preserves almost none of the original signal.

301 redirects are the correct choice for permanent URL changes because they pass the overwhelming majority of ranking signal to the destination URL over time. [VERIFIED]

Example redirect map format — every row should have a status code and a signal to verify, not just an old and new URL:

Old URL	New URL	Type	Ranking Signal to Verify
/product-category/running-shoes/	/collections/running-shoes/	301	Top 10 keywords, backlink count, indexed status
/blog/best-trail-runners-2025/	/guides/best-trail-runners-2025/	301	Featured snippet ownership, internal link equity
/about-us.html	/about/	301	Brand-query ranking, NAP consistency in schema
/checkout/step-2.php	(removed – no equivalent)	410 / noindex	Confirm zero organic sessions before removal

At launch, crawl the redirect map itself. Every old URL should resolve in a single hop to a 200-status new URL. Redirect chains (301 → 301 → 200) and loops are the most common silent failure in this step.

9. Enhance the New Site for SEO and AEO

A migration is the best opportunity a site will get to fix technical debt and add AEO infrastructure at once — templates are already being rebuilt, so schema, FAQ structure, and technical health improvements cost far less now than as a later retrofit.

Structured data

Rebuild schema per template type (Product, Article, FAQPage, Organization, BreadcrumbList) rather than copying the old theme’s markup verbatim
Validate every schema type with Google’s Rich Results Test after the new templates go live, not before

AEO / AI-citation readiness

Write FAQ answers as self-contained, directly quotable answers — the format most likely to be lifted into AI Overviews and chat answers
Keep any llms.txt file’s coverage complete, not just present — a partial llms.txt (e.g., missing a returns or policy page) can dilute citation the same way a missing page can
Preserve internal links between the new pages the way the old site’s link equity flowed — AEO surfaces still lean on the same crawlable link graph as classic search

Technical health and tracking

Re-run Core Web Vitals and crawlability checks (robots.txt, XML sitemap, canonical tags) against the new templates, not the old ones
Re-implement GA4 and conversion tracking on the new site before launch, and verify events fire correctly — not after the client asks why the dashboard is empty

10. Monitor After Launch — On a Fixed Schedule

Post-launch monitoring on a defined schedule catches ranking or indexing problems while they’re still cheap to fix. Waiting for a client to notice a traffic drop means the problem has already been live for weeks.

Day 1: confirm the new site is crawlable, the redirect map resolves cleanly, and GA4/GSC are receiving data
Day 7: check indexing status of the new URLs and any “discovered – not indexed” flags in GSC
Day 30: compare rankings for the top pre-migration keyword set against the new URLs; investigate any losses of more than a few positions
Day 90: compare organic sessions and conversions against the same period last year, adjusted for seasonality

Some ranking fluctuation in the first two to six weeks after a migration is expected while search engines re-crawl and re-evaluate the new URLs. A sustained drop past day 30 is a signal to investigate, not wait out.

[DISPUTED] The exact recovery window varies by site size, crawl frequency, and how many URLs changed — treat 30/60/90-day checkpoints as a monitoring cadence, not a guarantee.

Common Mistakes That Sink Migrations

Mistake	Why It Kills Rankings	Fix
Design starts before the SEO audit	New IA and templates get locked in before anyone knows which pages, links, or copy blocks carry ranking weight	Freeze design until the SEO blueprint is signed off (Step 1)
Redirect map built from a sitemap, not a crawl	Sitemaps miss orphaned pages, parameterized URLs, and pages that still rank but were removed from navigation	Build the map from a full Screaming Frog crawl plus GSC’s indexed-URL list
Copy rewritten without keeping the ranking terms	AI rewrites often smooth out the exact phrases and entities the page was already ranking for	Extract on-page terms and internal anchor text before rewriting; treat them as constraints, not suggestions
Redirects go live but nobody checks status codes	Chained or broken redirects (302s, redirect loops, 404s) leak link equity silently for weeks	Crawl the redirect map itself after launch; confirm every old URL resolves in one hop to a 200
Schema copied from the old theme without validation	Legacy schema often breaks on a new template’s markup, or duplicates entities	Re-validate every schema type with Google’s Rich Results Test after launch
No monitoring cadence after go-live	Ranking and indexing drops in week 2–4 get missed until a client asks why traffic is down	Set fixed checkpoints: Day 1, Day 7, Day 30, Day 90 in GSC and Analytics

A companion checklist to walk alongside this framework: The SEO Migration Checklist — a printable, phase-by-phase companion built to sit next to this exact framework. Every task maps to the step above it, and each one comes with a “Verify With” column so you know exactly which tool or report to check before ticking the box, plus space to assign an owner and date. Print it, work through it phase by phase, and don’t move to the next one until the current phase is signed off. Download the checklist →

The SEO Migration Playbook: How to Protect (and Grow) Rankings Through a Website Redesign