Content wins rankings, but on a Joomla site the technical layer decides whether that content is found once, under one URL, with the right title, and understood. This guide covers that layer: URLs, duplicates, titles and descriptions, structured data, sitemaps, multilingual tags and speed. It's what we check and fix on every Joomla site, including our own. We just went through our own blog with this list.
1. URLs: one page, one address
Joomla can reach the same article under many URLs: the menu route, /index.php/…, /component/content/article/…, with and without ?Itemid=, with and without a trailing slash. Search engines see each one as a separate page. The fixes are all in core:
| Setting | Where | Set to |
|---|---|---|
| Search Engine Friendly URLs | Global Configuration → Site | Yes |
| Use URL Rewriting | Global Configuration → Site | Yes (rename htaccess.txt to .htaccess, or the nginx equivalent) |
| Remove IDs from URLs | Articles → Options → Integration | Yes (check it on staging, see below) |
| Strict Routing | Plugins → System – SEF | Yes: redirects non-SEF URLs to the SEF one |
| Strict handling of index.php | Plugins → System – SEF | Yes: redirects /index.php/… away |
| Trailing slash for URLs | Plugins → System – SEF | one choice, enforced by redirect |
The SEF plugin options arrived in Joomla 5.1–5.2, and most sites never turned them on. They replace the hand-written redirect rules we used to put in .htaccess.
"Remove IDs from URLs" changes URLs on an existing site: /blog/12-my-post becomes /blog/my-post. Do it on a new site, or with 301 redirects from every old URL (the Redirect component or server rules). Turning it on without redirects is a reliable way to lose rankings.
# nginx – the Joomla front controller, with the URL rewriting it expects
location / {
try_files $uri $uri/ /index.php?$args;
}
2. Canonical tags: core only adds them in one case
Many people assume Joomla prints a canonical tag on every page. It doesn't. The System – SEF plugin only adds one when its Site Domain option is filled in. Components can add their own, but core content doesn't. Check your page source:
# terminal
curl -s https://example.com/blog/my-post | grep -io '<link[^>]*canonical[^>]*>'
If nothing comes back, set Site Domain, or add the canonical in your template the way we did on this site: path only, query string dropped except for pagination (?start=), and only on pages that are indexable.
3. Titles and descriptions
- Browser Page Title (article → Options → Layout, or the menu item's Page Display tab) lets the
<title>differ from the H1. We keep titles to 60 characters and put the long, descriptive version in the H1. One catch: Joomla only uses an article's Browser Page Title when a menu item links to that article directly. For articles reached through a category, you need a small template fix to apply it. - Site name in page titles (Global Configuration → Site): "after" is the usual choice, but a 55-character title plus " - Your Company" gets truncated. We drop the suffix when the result passes about 65 characters.
- Meta descriptions: Joomla uses the article's, then the menu item's, then the global one. A global description repeated on hundreds of pages is worse than none. Keep descriptions under 160 characters, and give category pages their own.
- Paginated lists (
?start=12) need a distinct title: "… - Page 2". Some SEO extensions append "Results from #12", which reads like a bug to users.
Find the gaps in bulk with SQL instead of opening articles one by one:
# terminal – published articles with a missing or too-long description, and long titles
mysql -e "SELECT id, CHAR_LENGTH(title) AS t, CHAR_LENGTH(metadesc) AS d, title
FROM jos_content WHERE state = 1
AND (metadesc = '' OR CHAR_LENGTH(metadesc) > 160 OR CHAR_LENGTH(title) > 60)
ORDER BY d;" mydb
4. Structured data: core does more than you think
Since Joomla 5, core ships schema.org support: a System – Schema.org plugin that prints a JSON-LD graph (Organization or Person, WebSite, WebPage), and schemaorg plugins per type: Article, BlogPosting, Book, Event, JobPosting, Organization, Person, Recipe and Custom. The article edit form gets a Schema tab where you pick the type. Set the base data (organisation name, logo) in the System – Schema.org plugin once.
Three things we fix on most sites:
- Choose BlogPosting for blog posts rather than the plain Article the core falls back to, with author, dates and an absolute image URL. Google ignores relative image paths in JSON-LD.
- Avoid duplicate BreadcrumbLists. The breadcrumbs module prints its own JSON-LD. If your template or an SEO extension prints another, the page has two competing trails.
- Validate after every template change with the Rich Results Test. A broken JSON-LD block fails silently.
# terminal – list the JSON-LD types a page outputs
curl -s https://example.com/blog/my-post \
| python3 -c "import sys,re,json; [print(json.loads(b).get('@type') or [n.get('@type') for n in json.loads(b).get('@graph',[])]) for b in re.findall(r'<script type=\"application/ld\+json\"[^>]*>(.*?)</script>', sys.stdin.read(), re.S)]"
5. Sitemaps and robots
- Joomla has no core XML sitemap. Use an extension that builds one from menus and content, and make sure it lists only canonical URLs: no
?Itemid=or/component/variants, no noindexed pages. robots.txt: Joomla's default no longer blocks/images/,/media/or/templates/, but sites upgraded from Joomla 3 often still have the old file. Blocking CSS and JS stops Google from rendering the page. Add the sitemap URL torobots.txt.- Noindex the utility pages: login, registration, search results, tag lists without content, checkout and account pages. Use the robots option of the menu item, or set it in the template for whole components.
- Empty categories that return a page with no posts are thin content. Unpublish them until they have posts.
6. Multilingual sites: hreflang from associations
With associations set up (Content → Multilingual Associations), the System – Language Filter plugin can print hreflang alternate tags. Turn on "Add Alternate Meta Tags", and use "Add x-default Meta Tag" for the fallback language. Hreflang only works when it's reciprocal. Every language version must link back. An article without an association in one language breaks the set for that pair, and the Multilingual Associations view shows you which ones are missing.
7. Speed is part of SEO
Core Web Vitals are a ranking signal, and most Joomla sites fail them for the same few reasons. The page cache is off, the hero image is lazy-loaded, too many scripts block rendering, and slow queries push the server time up. We cover each with code in Joomla performance from the code up.
8. Content structure that helps
Technical SEO gets the page found. Structure helps it rank:
- One H1 per page. Watch templates and modules that render module titles as
h1orh2above the content. - Internal links in the body of articles, between related posts and to service pages. Menus and "related articles" modules count less than a link in the text.
- Category pages with an introduction, their own title and description. They often rank for broader terms than any single article.
- Image alt text set per image, and covers sized to the layout.
The audit in one table
| Check | Healthy |
|---|---|
Non-SEF and /index.php/ URLs | 301 to the SEF URL |
| Canonical tag | on every indexable page, absolute, path only |
<title> | unique, ≤ 60–65 characters, "Page N" on pagination |
| Meta description | unique, ≤ 160 characters, also on categories |
| JSON-LD | BlogPosting/Article with absolute image; one BreadcrumbList |
| Sitemap | canonical URLs only, listed in robots.txt |
| Noindex | login, search, account, checkout, empty categories |
| hreflang | reciprocal, with x-default (multilingual sites) |
| Core Web Vitals | LCP image not lazy, page cache on for guests |
