If your business serves the Jordanian or Gulf market, your website needs to work in two languages. Your customers search in Arabic and English, often in the same session. A website that exists only in English misses Arabic search traffic entirely. A website that exists in both languages but is structured incorrectly confuses Google and may rank for neither.
I build bilingual websites and optimise existing bilingual sites as part of my SEO practice. The technical implementation matters more than most people realise — especially hreflang tags, which tell Google which language version to serve to which audience. Here is how to do it correctly.
Choose the Right URL Structure
There are three options for structuring a multilingual website. Only one is my recommendation.
Option 1: Subdirectories (Recommended)
English: domain.com/ — Arabic: domain.com/ar/
All content lives under one domain. Your domain authority consolidates rather than splitting across multiple properties. This is the structure I implement for every bilingual client.
Option 2: Subdomains (Not Recommended)
en.domain.com and ar.domain.com — Google treats subdomains as semi-separate entities. Your domain authority is partially split and SEO benefits are diluted.
Option 3: Separate Domains (Avoid)
Complete authority split. You are building SEO from scratch on two separate properties. Unless you have a strong strategic reason, avoid this.
The subdirectory approach wins because you build authority on one domain — every backlink, every page, every piece of content strengthens the whole site. Both language versions benefit from the total domain authority.
hreflang Tags: What They Are and Why They Matter
hreflang tags are HTML elements that tell Google which language and region each page targets, and which pages are equivalent translations of each other. Without them, Google guesses — and Google's guesses are often wrong.
I have audited bilingual sites where Google served the Arabic version to English-speaking users in the US and the English version to Arabic-speaking users in Jordan. The result: high bounce rates, poor user experience, and wasted SEO equity. All because hreflang was either missing or implemented incorrectly.
How to Implement hreflang Correctly
Every page that has an equivalent in another language needs hreflang tags in the <head> section:
<link rel="alternate" hreflang="en" href="https://domain.com/services/" /><link rel="alternate" hreflang="ar" href="https://domain.com/ar/services/" /><link rel="alternate" hreflang="x-default" href="https://domain.com/services/" />
The x-default tag tells Google which version to show when the user's language does not match either version. I always set x-default to the English version as the fallback.
The Rules That Must Be Followed
- Every page must reference itself. The English page must include an hreflang tag pointing to itself AND to its Arabic equivalent. The Arabic page must do the same in reverse. If the references are not bidirectional, Google ignores them.
- URLs must be absolute. Use the full URL including
https://, not relative paths. - Every language page must be self-canonicalising. The Arabic page's canonical tag should point to itself (the Arabic URL), not to the English version. A common mistake is having all canonical tags point to the English version — this tells Google to ignore the Arabic pages entirely.
- hreflang must be on every page pair. Not just the homepage. Every English page that has an Arabic equivalent needs the tags, and vice versa.
Arabic Content: Not Translation
The Arabic version of your website should not be a word-for-word translation of the English version. Arabic search behaviour is different from English search behaviour. The keywords people use in Arabic do not map one-to-one to English keywords.
Arabic keyword research is a separate process. I research what people actually search for in Arabic using tools like Ahrefs and Google Search Console's Arabic query data. The content structure, page titles, meta descriptions, and body content are all written for Arabic search intent — not translated from English.
Machine translation produces Arabic content that native speakers immediately recognise as translated. It is grammatically awkward, misses cultural nuances, and often uses formal MSA phrasing when colloquial or regional phrasing would be more natural. Native Arabic content is not optional — it is the standard.
RTL Considerations for SEO
- Page titles and meta descriptions in Arabic must read naturally in RTL. Mixed Arabic-English titles can render awkwardly in search results if bidirectional text handling is not correct.
- Structured data for Arabic pages should contain Arabic text — your business name in Arabic, service descriptions in Arabic, and Arabic-language FAQ content.
- URL slugs should be descriptive in English — Arabic characters in URLs are technically supported but can cause display issues. I use descriptive English slugs for Arabic pages:
domain.com/ar/seo-services/rather than Arabic characters in the URL.
Common Bilingual SEO Mistakes
- No hreflang at all. The most common mistake. Two language versions exist but Google has no instructions on which to serve to whom.
- Canonical tags pointing all Arabic pages to English. This de-indexes your Arabic content. Each page should self-canonicalise.
- Partial translation. Some pages have Arabic versions, others do not. The site structure is inconsistent, confusing both users and crawlers.
- Ignoring Arabic search volume. In many industries, Arabic search volume exceeds English — especially for local services in Jordan and the Gulf.
For the full guide on all aspects of SEO, read my Complete SEO Guide for Businesses in Jordan.
13 years of documented results across Jordan and the Gulf. Explore the full SEO service or book a free strategy call.