WordPress SEO in 2026: Fix the Defaults
Say you inherit a 900-page WordPress site from a mid-sized retailer in Manchester. Traffic has slid for nine months. The team blames the last core update. Then you open Search Console and see something else. Google has indexed 4,100 URLs. The site has 900 real pages.
The other 3,200 are tag archives, author archives, date archives, and attachment pages. Nobody wrote them. Nobody signed them off. WordPress built them on its own, the moment a writer hit Publish.
That gap is the whole story of WordPress SEO. The platform is generous. It gives you clean markup, readable URLs, a sitemap, and an editor your writers will actually open. It also ships defaults built for a 2005 personal blog, not a 2026 commercial site. Most teams never touch them.
This guide covers the second half. Not what a title tag is. What WordPress does with yours, where the platform helps you, where it works against you, and which changes are worth an afternoon of your time.
What makes WordPress SEO different from SEO on other platforms?
WordPress rarely blocks good SEO. It just does a lot of things for you without asking, and those automatic decisions become your indexation, your page speed, and your crawl budget. Platform-specific work means auditing what WordPress publishes on its own, then deciding what stays. That is a different job from a generic checklist.
Here is the promise. By the end you will know which permalink structure to pick, which auto-generated archives to remove from the index, which handful of plugin settings earn their keep, and how much your page builder costs you in Core Web Vitals. You will also know what to skip. If you want the platform-neutral basics first, start with our on-page SEO checklist and come back here.
Three positions run through this piece, and two of them are unpopular. First, most SEO plugin settings are noise. Six or seven matter, and the rest exist because a rival shipped them. Second, the platform's largest liability is publication without permission, not slow PHP. Third, your host and your caching layer decide more of your Core Web Vitals score than your theme does.
One objection first. Yes, plenty of huge WordPress sites rank fine with default settings. They rank in spite of them, usually because they have links and a brand. If you do not, the defaults cost you. Good WordPress SEO on a mature site is mostly subtraction.
What should your permalink structure be?
Set permalinks to Post name and leave them alone. The default plain structure produces query-string URLs that read like machine output and carry no keywords. Category-based URLs look tidy on day one and turn into a redirect project the first time you reorganize your taxonomy. Post name wins because it is short, stable, and boring.
Two caveats matter. If the site already ranks on dated URLs, changing the structure means a full redirect map, and that is a real project rather than a checkbox. Read our guide to redirects and site migrations before you touch a live structure. Second, WooCommerce adds its own product and shop base, and those bases behave like a second permalink system. Decide early whether product URLs sit at the root or under a shop path, because moving them later hits your best commercial pages. The WordPress permalink documentation lists every available tag.
Why does WordPress index pages nobody wrote?
WordPress creates an archive for every tag, every category, every author, and every publication date, plus an attachment page for every uploaded file. Each one is a live, indexable URL unless you or a plugin turns it off. On a busy site that turns a few hundred articles into several thousand thin pages competing with the articles themselves. Attachment pages are the one exception worth checking first, because modern SEO plugins switch them off and newer WordPress versions do the same on fresh installs. You mostly find them still live on sites that have been running for years.
This is the strongest opinion in the piece. The biggest WordPress SEO liability is not slow code. It is publication without permission. No writer asked for an author archive at slash author slash admin. The platform shipped it anyway, and Google now spends crawl effort on it.
A working rule for each archive type
My rule is blunt. Keep category archives if a human would ever browse them, and only if you write real copy at the top. Remove tag archives from the index unless tags do genuine navigation work, which on most sites they do not. Remove author archives on any site with one or two writers. Remove date archives always. Redirect attachment pages out of the index, which most SEO plugins now do by default, either to the parent post or to the file itself, though you should confirm rather than assume.
Then check the result. Pull the indexed count in Search Console, compare it with your real page count, and work the gap down. Our technical SEO audit process covers the crawl side in more depth, and our content pruning and refresh guide handles the thin pages you did write.
Which WordPress SEO plugin settings actually matter?
Roughly seven settings do real work. Everything else in a modern SEO plugin is either a preference, a duplicate of something WordPress already handles, or a feature built to match a rival's list. Yoast, Rank Math, SEOPress and All in One SEO all ship those seven. Configure them, then close the settings screen and go write something.
| Setting | Why it earns its place |
|---|---|
| Index rules for tag, author and date archives | Stops the thin-page problem at the source |
| Title and description templates per post type | Fixes thousands of pages in one edit |
| Canonical handling | Keeps paginated and filtered URLs from splitting signals |
| Sitemap contents | Tells Google which post types you treat as real |
| Organization or person schema | Feeds the knowledge panel and AI answers |
| Breadcrumbs | Only if your theme does not already output them |
| Attachment page redirects | Removes a whole class of empty URLs |
What can you ignore? Most of the rest. The social preview tab is a convenience, not a ranking feature. Our meta tag generator is a faster way to draft a clean set of meta tags before you touch any plugin settings. The readability traffic lights measure sentence length, not rankings. Keyword density scoring encourages exactly the writing Google has punished for a decade. For structured data beyond the basics, our schema markup and JSON-LD guide takes you past what a plugin wizard covers, and our word counter settles length arguments faster than a sidebar panel.
Should you use the built-in sitemap or a plugin sitemap?
Use the plugin sitemap on any site with archives you want excluded. WordPress core has shipped an XML sitemap at wp-sitemap.xml since version 5.5, and it works, but it lists every public post type and taxonomy by default. Changing that needs code, and a plugin gives you checkboxes instead. Small brochure sites can happily keep the core version.
Two things go wrong here. Teams end up with two sitemaps making different claims. Most SEO plugins now switch the core sitemap off when you enable theirs, so this is rarer than it was, but confirm it rather than assume. Pick one and turn the other off. Teams also submit a sitemap once and never look at the coverage report again, which is where the real signal lives. Our XML sitemaps and indexing guide walks through that loop.
Do page builders really hurt your Core Web Vitals?
Yes, most of them do, and the damage is measurable. Elementor, Divi and similar builders wrap content in extra nested markup and ship their own CSS and JavaScript on every page. That inflates Largest Contentful Paint and Interaction to Next Paint. Cumulative Layout Shift usually suffers too. Gutenberg, the block editor, is far lighter because it saves plain HTML.
Be fair about the trade-off. A builder lets a marketing team ship landing pages without a developer, and that speed has real commercial value. Ripping one out of a 200-page site is a rebuild, not a fix. So the honest question is narrower. Does the builder sit on your money pages, or only on campaign pages nobody links to?
If it sits on your money pages, do three things. Turn off the builder's global assets on templates that never use it. Load its icon fonts and animation libraries only where needed. Then judge the result on the field data in the Chrome User Experience Report rather than the lab score in PageSpeed Insights. A lab run is one simulated device on a throttled connection, and it cannot measure Interaction to Next Paint at all. Google grades you on what real visitors experience. The Core Web Vitals reference on web.dev and our own page experience guide cover the thresholds.
How much do extra plugins and cheap hosting really cost?
Plugin count matters less than plugin behavior. Twenty quiet plugins can beat five noisy ones. What hurts is the plugin that queries the database on every request, adds a script to every page, or fires a scheduled task that WordPress runs on a real visitor's page load. That last one is wp-cron, and it surprises people.
Hosting is a WordPress SEO decision, and I will defend that. Time to first byte sets the ceiling on how fast your Largest Contentful Paint can ever be. A shared plan in a data center 4,000 miles from your buyers starts you hundreds of milliseconds behind. Managed hosting, a Redis object cache, and a full-page cache in front of PHP move that number more than any theme tweak. Speed aside, it is worth running a key page through our SEO report for an outside view of how it is put together.
When a CDN is not worth it
A CDN like Cloudflare helps most when your audience spans the United States and Europe. If every customer sits in the Netherlands and your server sits in Amsterdam, a CDN buys you offload and security rather than speed, and you should be honest about which one you are paying for. This is one of the few WordPress SEO calls where geography decides the answer, not configuration. Say so out loud instead of adding another layer.
What does the media library do to your images?
It makes copies. Every upload generates several resized files, and your theme often adds more, so one photo can become six files on disk. WordPress then serves a responsive image set, which is genuinely helpful. The problem is the source file. WordPress re-compresses the sizes it generates and scales very large uploads down to a smaller serving copy, but it will not change your format, and a bloated source still produces bloated copies. A 4 MB hero photo can reach the page far heavier than it needs to be.
Do three things and stop. Convert new uploads to WebP or AVIF, either at the host level or with one image plugin. Set sensible dimensions before upload, because editors will drag in whatever the photographer sent. Then confirm your hero image is not lazy loaded, since recent WordPress versions try to handle that automatically and page builders sometimes undo it. That single mistake wrecks Largest Contentful Paint more often than heavy file sizes do. Our image optimization guide and our Core Web Vitals guide for 2026 have the detail.
How should European teams handle multilingual WordPress?
Pick one approach and commit. WPML and Polylang both run translations inside a single install, which keeps admin simple and hreflang consistent. Multisite gives each language its own site and its own risk of drift. For a team selling across Germany, France and Spain, the single-install route usually wins on maintenance.
Get hreflang right or skip the effort. Every language version must point at every other one, including itself, and the tags must agree in both directions. A half-built set is worse than none. Watch for machine-translated pages that nobody reviews, because thin translated copy fails the same bar as thin English copy. Run local-language keyword research per market rather than translating your English terms. One more European detail: cookie banners built for GDPR often delay rendering, and then run heavy scripts the moment a visitor clicks accept, which hurts Largest Contentful Paint and Interaction to Next Paint on the pages you paid to promote.
What should you ignore in WordPress SEO?
Ignore the perfect Lighthouse score, the green light in your plugin sidebar, and the panic about wp-json. Those three eat more agency hours than they return. Spend the time on indexation, internal links, and the pages that already earn money. Each one feels productive, which is exactly why it survives.
The REST API at wp-json exposes user names and post data on many installs. That is a security and privacy question, not a ranking one. Google does not demote you for it. Lock it down as good hygiene, then move on. Crawler control is a different matter, and our AI crawlers and robots.txt guide covers the rules worth writing.
Skip the 100 score too. A site at 92 with real content beats a site at 100 with nothing to say. And if you run a small business site with twelve pages, skip most of this article. Set permalinks to Post name, remove date and author archives, compress your images, and go do local SEO and Google Business Profile work instead. That is genuinely all you need. The Google Search Central starter guide covers the rest.
Frequently asked questions about WordPress SEO
Is WordPress good for SEO?
Yes, WordPress is good for SEO out of the box, and better once you change a few defaults. It gives you clean HTML, editable titles and descriptions, a working sitemap, and a huge plugin ecosystem. The catch is what it publishes automatically. Tag, author, date and attachment pages all go live without a decision from you. Fix those, pick sensible permalinks, and WordPress stops being a platform problem and becomes a content problem.
Which SEO plugin is best, Yoast or Rank Math?
Both are fine, and the choice matters less than people claim. Rank Math packs more features into the free tier and has a lighter admin. Yoast has a longer support history and cleaner documentation. All in One SEO and SEOPress do the same core job. Pick the one your team will actually configure, then use about seven of its settings. Switching plugins for a feature list is rarely worth the migration risk.
How many plugins are too many for SEO?
There is no magic number. A site with thirty well-built plugins can outrun a site with eight badly-built ones. Judge by behavior instead. Check how many database queries each plugin adds, whether it loads scripts on pages that never use it, and whether it hooks into every request. Two or three heavy plugins cause most speed problems. Audit what each one loads before you count anything.
Should I delete tag pages on WordPress?
Do not delete them, remove them from the index. Deleting tags breaks internal links and strips context from posts. Instead, set your SEO plugin to mark tag archives as noindex, and let them keep working as navigation for readers. If a tag archive genuinely earns search traffic, and a few do on large publisher sites, keep that one indexed and write an introduction for it. Check Search Console before you decide.
Does the WordPress block editor help page speed?
Yes, compared with most page builders. Gutenberg saves posts as plain HTML with simple classes, so the output stays close to what you wrote. Elementor and Divi store their own structures and load extra CSS and JavaScript on every page. That difference shows up in Largest Contentful Paint and Interaction to Next Paint. The block editor is not free of bloat, though, because block libraries and heavy themes add their own weight.
Do I need a sitemap plugin if WordPress has one built in?
Only if you need control. The built-in sitemap lists every public post type and taxonomy, and changing that requires code. An SEO plugin gives you checkboxes for the same job, which matters once you noindex archives. Whichever you use, disable the other one so Google is not reading two conflicting files. Small sites with no unwanted archives can stay on the core sitemap without any downside at all.
How do I fix Core Web Vitals on a WordPress site?
Start with the largest image on the page. Make sure your hero image loads eagerly, carries a high fetch priority hint, and ships as WebP or AVIF at sensible dimensions. Next, cut render-blocking CSS and JavaScript from plugins you do not use on that template. Then add a full-page cache and a decent host, because time to first byte caps everything else. Measure with field data, not one lab test.
Is wp-json a security or an SEO risk?
It is a privacy and security question, not an SEO one. The REST API at wp-json can expose user names and content on default installs, which helps attackers guess logins. Google does not rank you lower for leaving it open. Restrict the user endpoint, keep WordPress and PHP updated, and use strong credentials. If you disable the API entirely, test the admin first, because the block editor and many plugins depend on it.
What is the best permalink structure for WordPress?
Post name is the right default for almost every site. It gives short, readable, keyword-bearing URLs that survive a redesign. Date-based structures make URLs look stale and force a redirect project every time you refresh content. Category-based structures break the moment you reorganize your taxonomy. If you already rank on a different structure, do not change it without a full redirect map and a plan to monitor errors afterwards.
Should a WooCommerce store handle SEO differently?
Yes, in a few specific ways. Product variations, filters and sort parameters generate large numbers of near-duplicate URLs, so canonical tags and crawl rules matter far more than on a blog. Product schema needs accurate price, currency and stock data for European and US listings alike. Category pages usually deserve real copy, because they target the broad commercial terms. Everything else follows the same rules as any WordPress site.
Where to start on Monday morning
Go back to that Manchester retailer. Nothing in this list needs new content. The tag, author and date archives leave the index. Attachment pages redirect to their parents. The oversized hero images shrink. The indexed count starts moving toward the real page count, and the pages that actually sell get room to breathe.
Priority order, if you only get one afternoon. First, fix indexation of the auto-generated archives, because it is the largest and cheapest win. Second, check the hero image on your top ten pages. Third, set title and description templates per post type. Then build proper content clusters and pillar pages on top of the clean structure.
Here is my prediction. As AI answers take more of the click, WordPress SEO will reward sites with clean, small, unambiguous URL sets even more than it does now. Machines do not enjoy guessing which of your four near-identical URLs is the real one. Neither do people.
What is the strangest thing WordPress has published on your site without asking?