The Migration Equity Recovery Audit: Finding the Links Your Replatform Left Behind

A platform migration is usually framed internally as a marketing win: a faster CMS, cleaner templates, a better editorial workflow. The link equity consequences rarely make it onto the launch deck. Within a quarter, though, US teams often discover that a handful of high-value URLs — the ones that took three years of digital PR to earn — now resolve to a generic category page instead of the page that attracted the citation.

That is the replatform leak. It is seldom one catastrophic error. It is dozens of small mapping decisions made under deadline pressure, each of which quietly steers a referring URL away from its intended destination.

Why the damage stays invisible for months

301 redirects are forgiving to humans and unforgiving to crawlers. A browser follows a five-hop chain instantly. A crawler registers a chain, defers the pass, and frequently treats the terminal destination as a soft 404. The documentation is explicit that long chains and redirect loops degrade signal transfer, yet the Search Console report that would reveal the issue — Page with redirect — does not distinguish between a single clean hop and a six-step cascade routed through two legacy domains.

Compounding matters, rankings rarely collapse on day one. The decay curve typically trails the cutover by six to twelve weeks, arriving just after the migration war room has disbanded and the team has moved to the next quarter's roadmap. By then the cause is no longer obvious to anyone who was not in the room.

Phase one: reconstruct the legacy URL inventory

An audit is only as good as its enumeration. Before touching a single redirect rule, assemble a complete inventory of every URL that existed in the twelve months preceding launch. Cross-reference four sources:

• Historical sitemaps. Pull every XML sitemap served over the prior year, not merely the final version before cutover.
• Wayback Machine CDX API. Query the legacy host for all captured paths; this surfaces pages that never appeared in a sitemap, including campaign and landing pages built outside the CMS.
• Search Console, both properties. Keep the legacy property verified after launch. Its Links report remains the most reliable record of which legacy URLs actually attracted external citations.
• Server access logs. The 45 days before cutover show which paths crawlers and referrers were hitting most often, which helps weight the rest of the list.

Deduplicate the combined list, normalize trailing slashes, lowercase, and strip non-essential query parameters. For a mid-size US ecommerce or B2B publisher, expect somewhere between 8,000 and 40,000 unique paths.

Phase two: trace every referring URL to its terminal status

Match the inventory against inbound-link exports, then crawl each mapped destination with a tool that reports hop count and terminal status code. Flag four conditions, ordered by severity:

1. Unmapped 404s. Legacy URLs with external links and no redirect rule at all. These are pure loss.
2. Chains of three or more hops. Often the byproduct of layered plugins, CDN rules, and application-level redirects that no single owner controls. Flatten every chain to a single hop where the destination is stable.
3. Redirects to the homepage. The most common migration shortcut and the most damaging. It tells the crawler the old page is gone while offering no topical substitute, which is a poor signal for both the referring page's subject matter and the user's intent.
4. Cross-domain chains. Domain consolidation projects frequently route through an intermediate property, doubling hops and dividing signals between hosts.

Record the referring domain, the anchor text, the original target, the current terminal URL, and the hop count in a single working sheet. That sheet becomes the backlog.

Phase three: identify orphaned legacy paths

Some pages had no obvious successor, and the temptation is to sweep them into the homepage rule. A better pattern is to assign each to the nearest topical equivalent, judged by the query the page was built to answer rather than by its URL structure. Three resolutions are defensible:

• A direct equivalent that serves the same intent.
• A curated hub page that genuinely answers the original query and links onward to related content.
• A 410 for material retired with no replacement. An honest 410 is preferable to a misleading 301, because it stops the crawler from re-evaluating an irrelevant destination.

Reserve the homepage redirect exclusively for genuinely brand-level pages such as press mentions of the company itself.

Phase four: audit internal link pathways

External equity draws the attention; internal pathways determine whether that equity ever reaches commercial pages. Three checks matter most. First, navigation and footer truncation: a migration that trims the main nav from nine items to six removes internal links from every page on the site at once. Second, breadcrumb loss, which severs the shortest path from deep content to category hubs. Third, crawl the live site for internal links still pointing at legacy paths — each one burns an extra hop and slows discovery of new URLs.

Ranking recovery work by estimated authority impact

A long backlog needs a defensible order. A workable priority score multiplies the referring domain's authority tier (1 to 3), the topical relevance of the referring page to the target (1 to 3), and a link-type weight (editorial body link above resource page above directory listing), then divides by the hop count plus one. Sort into three tiers:

• Tier 1: external links to commercial-intent legacy URLs that currently 404 or travel three or more hops.
• Tier 2: high-authority editorial citations landing on redirects to the homepage.
• Tier 3: long-tail volume — many low-authority links to retired blog paths.

Tier 1 usually represents under 5 percent of the backlog and the majority of the recoverable value. Fix it first, verify with a re-crawl within 72 hours, and log the change against impressions on the destination.

A 30/60/90 cadence that keeps the leak closed

In the first 30 days, flatten all chains to one hop, resolve 404s, and restore internal links. Between days 30 and 60, replace homepage fallbacks with topical equivalents and re-submit the legacy sitemap for crawl. From day 60 to 90, compare Search Console link data on the new property against the pre-migration baseline and reconcile any referring domain that still points somewhere unintended. Outreach to publishers is worth considering only after the technical layer is clean; asking for a link update on a URL that still redirects poorly wastes the request.

The lesson is structural rather than technical. Treat the redirect map as an asset with an owner, a version history, and a quarterly review, and the next replatform will cost far less equity than the last one did.