Rank Math Expert
Beyond the 100/100 Score

Rank Math is a good plugin, it is free, and it is responsible for a specific and widespread misunderstanding: that a green 100/100 in the editor sidebar means a page is optimised. It does not. The score is a checklist that runs in your browser, and Google has no access to it, no knowledge of it, and no interest in it.

What the checklist measures is hygiene. Keyword in the title, keyword in the URL, keyword in the first paragraph, a subheading, an image with alt text. All worth doing and none of it a reason to rank. A thin 400-word page can score 100. A genuinely authoritative page can score 68 because the keyword is not in the slug. Chasing the number instead of the outcome is the most common mistake I see on otherwise well-run sites.

A Rank Math expert is useful for the settings three menus deeper, which do change what Google sees. Sitemap Settings decides which content ever gets submitted, and I regularly find entire custom post types live, internally linked, and completely absent. The schema module decides your entity graph, and duplicate Organization nodes from a plugin and a theme both claiming to be the business is an unhelpful signal. Redirections decides whether a migration kept its rankings.

The distinction that matters is between what the plugin asserts and what Google verifies. Google’s structured data guidelines are explicit that markup must match visible content. Schema claiming things the page does not show gets discounted, and a plugin will happily let you claim them.

Start with the sitemap, not the score

The sitemap is the closest thing to a statement of what you believe deserves to rank, and most of them are wrong by accident. Custom post types default to off. Individual posts get noindexed during a cleanup and never restored. Thin archive pages get submitted because nobody turned them off.

So the first pass compares what is published against what is submitted, in both directions. Pages live but missing, and pages submitted but not worth an index slot. That single comparison usually surfaces more than the rest of the audit combined.

One source owns the schema graph

Duplicate structured data is the quietest technical SEO problem there is, because nothing errors. Rank Math emits an Organization node. A custom theme emits its own. Both are valid, both describe the same business, and they disagree on an address or a rating.

The fix is deciding which one owns the graph and switching the other off, then validating against both the Rich Results Test and the Schema.org validator, because they catch different things. On template-driven sites the theme usually wins, since it can express relationships a plugin cannot. Relevant on any custom build.

Turn off what costs more than it returns

Rank Math ships with modules enabled that most sites never use, and several carry real cost. The Analytics module writes Search Console data into custom database tables that grow indefinitely, which on shared hosting makes wp-admin measurably slower for data you already have in Search Console for free.

The same logic applies to Content AI, video and news sitemaps, and the bbPress and BuddyPress integrations. Every enabled module is code loading on every request. Turning off what you do not use is the cheapest performance work available, which pairs with proper speed optimization.

Migrate with a crawl, not with hope

Yoast to Rank Math looks like a button and is not. The importer handles standard post meta well and regularly misses custom post types, taxonomy meta, and anything a third-party plugin stored through Yoast’s filters.

So the sequence is a database backup, a full crawl capturing every title, description and canonical before the switch, then the import, then the same crawl compared against the first. Anything that changed gets fixed by hand. Without that comparison, a site can lose meta on a third of its URLs and nobody notices until traffic drops a month later. Same discipline applies to a redesign or a maintenance plan that keeps it honest.