An SEO plugin can report a successful import while a product publishes a different title, a category loses its description, or an article points to the wrong canonical URL. Treat the switch as a comparison of published pages, not just settings screens. The aim is to preserve the metadata you intended for important URLs and to identify any deliberate changes. No migration checklist can guarantee unchanged search rankings.
Save a baseline before changing plugins
Back up the WordPress database and site files, and confirm that the backup can be restored. Decide who can authorise a rollback. Both Yoast’s Rank Math migration instructions and Rank Math’s Yoast migration instructions recommend a backup before importing. Keep it until the new plugin's output has passed your checks. The backup protects stored data; it does not show what a visitor or crawler received before the switch.
Choose URLs that expose different behaviour. Include a product, a product category, an article and a page that relies on default title or description templates. Add pages with manually entered metadata, a custom canonical, a noindex rule, or a required redirect. If the shop uses other product types or taxonomies with separate settings, sample those too. Record exact URLs in one worksheet and check those same addresses afterwards.
Give the sample enough variety to reveal a shared setting that failed. Include a product with a hand-written title and one that uses the default. Choose a category whose description was edited, plus one that was left alone. If the shop has several languages or a separate mobile address, include the relevant versions only when they are part of the same migration. Do not replace a failed sample URL with an easier page after the import.
For each page, save its visible heading, the values entered in the current plugin, and the relevant published HTML: page title, meta description, canonical link, robots directive and structured data. Note any X-Robots-Tag response header. Record the response status and final address when a URL redirects. Write down the XML sitemap location and whether each sampled canonical URL appears there. Mark each title and description as manually entered or generated by a template; otherwise, a changed default can be mistaken for a failed import.
Keep redirects and structured data in the inventory even if another tool controls them. You need to verify their output, but they may have no field for the SEO plugin to import. Record who owns each rule so that a missing redirect is investigated in the right place.
Use the import path for the direction of travel
Follow the destination plugin's current instructions rather than assuming that its controls match the plugin being replaced. Yoast describes deactivating Rank Math, importing through Yoast's tools, checking imported fields, completing its first-time configuration and then cleaning up old data. Rank Math describes importing Yoast data through its setup wizard or Import & Export area. These documented paths differ; choose the one that matches your move.
Schedule the change for a time when you can inspect the site immediately. Note the plugin versions, import options and any error or completion message. If the import offers separate choices for settings and post metadata, record which ones you selected. Retain the backup and old data until you have compared public output. An import-complete message confirms that a process ran, not that every page type now publishes the intended fields.
Open the sampled products, categories and articles in the destination plugin. Compare editable fields with your baseline, including intentional blanks. A blank title may use a template; a newly applied template can change many pages even when custom titles migrate correctly. Check whether the importer covered each relevant post type, taxonomy and exception. Record fields with no obvious destination rather than counting them as imported.
Check the HTML and responses visitors receive
Request each sampled URL and compare it with the saved baseline. Check its response status and final address, then inspect the HTML for the intended title, description, canonical link, robots directive and structured data. Look for duplicate or conflicting tags as well as missing ones. A familiar value in the editor is insufficient if the published page differs. If a cache is involved, clear or bypass it as appropriate and repeat the request before diagnosing the plugin.
Work through page types separately. A product with custom metadata tests individual field transfer. A category tests taxonomy metadata. An article tests ordinary post fields. A page without custom fields tests the destination's default templates. For an exception page, compare the exact canonical, robots rule or redirect that made it an exception. If several pages of one type differ in the same way, inspect that type's templates and settings before repairing pages individually.
Check the signals outside the HTML fields too. Request old URLs that should redirect and confirm their status and destination. Inspect the sitemap at its recorded address and check whether the sampled canonical URLs are present where intended. Compare any X-Robots-Tag header with the page's robots meta directive. A page that loads normally may still send conflicting indexing or canonical signals.
Review structured data by page type rather than expecting the same markup everywhere. Compare a product with the product baseline and an article with the article baseline. Note missing or changed fields, but check whether the theme, a commerce extension or another plugin produced them before assigning the fault to the SEO import. Where two components produce the same type of markup, investigate the duplicate output instead of deleting one blindly.
Google treats redirects and canonical annotations as strong canonical signals, while sitemap inclusion is a weaker signal. It can still choose a different canonical from the one you specify. Use Google’s canonicalization guidance when reviewing a discrepancy. The migration check establishes which signals your site publishes; it cannot prove which URL Google will select.
Do not judge title migration by the wording shown in search results. Google may form a title link from the HTML title, visible headings, prominent text and other signals, as its title-link documentation explains. Compare saved and current page HTML first. Search appearance can be observed later, but it is a different test from whether the plugin transferred a title field.
Repair gaps or use the planned rollback
Set failure triggers before the import. Examples are missing titles or descriptions across a page type, unexpected noindex directives on pages intended for indexing, changed canonical targets, broken required redirects, or structured data disappearing from pages where it was present. A single exception may be fixed directly; a repeated fault across products or categories warrants a broader check. Agree who decides whether to pause or restore.
When a result differs, check whether the cause is an import mapping, a new template, another plugin or theme, or cached output. Fix the cause and repeat the same URL comparison. If the agreed trigger is met and a prompt repair is uncertain, follow the rollback procedure and verify the restored public pages. An intentional improvement may differ from the baseline; record its reason instead of forcing the old output back.
Close the migration record with the sampled URLs, before-and-after fields, public responses, repairs and unresolved mappings. Keep the record with the backup location and rollback decision. That gives the next reviewer a concrete account of what changed and which page types were actually checked.