Adding Arabic to a UAE business website does one of two things. It roughly doubles the audience you can reach, or it splits your existing rankings across two versions of the same page that compete with each other and neither wins. The outcome is decided almost entirely by four technical decisions made before anybody translates a single word. This article covers those four decisions, in the order you have to make them. If you have not yet decided whether a second language is worth it at all, start with our piece on that question and come back here for the how.
First: should your UAE business have an Arabic website at all?Decision One: Where the Second Language Lives
This is the decision that is expensive to change later, so make it first and make it deliberately. There are three options.
- Subdirectory. yoursite.com/ar/ sitting inside your existing domain. All the authority you have built stays in one place and benefits both languages. Simplest to manage, cheapest to host, and the right answer for the overwhelming majority of UAE businesses.
- Subdomain. ar.yoursite.com. Search engines treat subdomains as partially separate properties, so authority does not flow as freely. Occasionally justified when the two versions are genuinely different sites run by different teams. Usually it is chosen by accident because a plugin defaulted to it.
- Separate country domain. yoursite.ae alongside yoursite.com. This is a country signal, not a language signal, and it means building authority twice from zero. Worth it if you are running genuinely separate operations in different countries. Almost never worth it for one UAE business wanting a second language.
Default to the subdirectory unless you have a specific reason not to. The predictable regret is the business that put Arabic on a subdomain three years ago and now has two sites, one of which nobody has updated since launch.
Whichever you choose, do not detect the visitor's language and redirect them automatically. It seems helpful and it causes two problems. Googlebot crawls from a fixed location and may only ever see one version, and a bilingual reader who wants the English page gets forced into Arabic every time. Offer a clearly visible language switcher instead, remember the choice, and let people decide.
Decision Two: Hreflang, and the One Mistake Everyone Makes
Hreflang tags tell Google that two pages are the same content in different languages, so it can serve the right one to the right person instead of picking a favourite and burying the other.
Four rules cover nearly every case.
- Use full codes, not bare language codes. en-AE and ar-AE rather than en and ar. If you also serve Saudi Arabia, ar-SA is a separate entry and worth having, because search behaviour and terminology genuinely differ between the two markets.
- Make the tags reciprocal. Every version must point to every other version and to itself. This is the mistake almost everyone makes: the English page lists both versions, the Arabic page lists only itself or nothing at all, and Google discards the entire relationship. One-way hreflang is the same as no hreflang.
- Set an x-default. This is where visitors go when they match none of your language versions. Usually your English homepage.
- Point at the exact URL, including the protocol and any trailing slash, and make sure each target actually returns a 200 status rather than a redirect.
You can place these in the page head, in your sitemap, or in HTTP headers. Pick one method and use it consistently. Mixing two methods across a site is a reliable way to produce conflicting signals that are difficult to debug later.
Hreflang does not consolidate authority the way a canonical tag does. It tells Google these pages are equivalent alternatives, not that one supersedes the other. Never point a canonical tag from your Arabic page to your English page. That tells Google the Arabic version should not be indexed at all, which is the opposite of why you built it. It is an easy mistake to make, because pointing a canonical at the 'main' version feels intuitive and is exactly wrong here.
Decision Three: How the Translation Actually Happens
There are three approaches and the middle one is where almost everyone should land.
Pure machine translation, meaning a plugin that translates on the fly with nobody reviewing it, is the cheapest and the worst. Google treats unreviewed machine translation as scaled content, so it does not rank. More importantly, the readers who do arrive can tell immediately, and in a market where trust drives B2B enquiries, an obviously machine-translated page costs you more credibility than having no Arabic page at all.
Machine translation with human editing is the practical answer for most businesses. Modern engines get you 80 percent of the way on straightforward content, and a native speaker fixes terminology, tone and the parts that came out nonsense. Budget for the editing as a real line item, not an afterthought, and make sure the editor knows your industry vocabulary. Technical product terms are where machine translation fails hardest and where a wrong word costs you a sale.
Full human transcreation, where a writer produces the Arabic version natively rather than translating, is right for your homepage, your top three service pages, and anything designed to persuade. It is overkill for a product specification table.
Decision Four: Which Pages, and in What Order
Do not translate the whole site. This is the mistake that turns a good idea into an abandoned project: 140 pages get quoted, the price comes back, the whole thing gets shelved, and two years later there is still no Arabic version.
- Homepage. Non-negotiable, and worth doing as transcreation rather than translation.
- Your top three to five service or product pages, chosen by which ones currently generate enquiries rather than which ones you like most.
- Contact page, including your address written the way an Arabic speaker would search for it.
- The two or three pages that answer the questions customers actually ask before buying.
That is usually eight to twelve pages. Launch those, give it three to six months, and look at the data before committing to anything else. If the Arabic pages bring enquiries, expand deliberately. If they bring nothing, you have learned that cheaply rather than expensively.
Blog content translates worst of all, incidentally. It is high volume, low intent, and ages fast. Translate service pages first and blog posts last, if ever.
What Arabic SEO Is Not
It is not your English keywords rendered in Arabic script. Two things break that assumption.
First, search terms genuinely differ. People searching in Arabic use different phrasing, and a direct translation of your best English keyword is often not what anybody types. Keyword research has to be done again, in Arabic, by someone who speaks it. Translating your keyword list is not keyword research.
Second, a large share of UAE search happens in a mix. Arabic speakers frequently search for technical terms, brand names and product categories in English, or in Arabic with the English term transliterated. Your Arabic content should reflect how people actually search, which sometimes means keeping the English term in place rather than translating it into a word nobody uses.
And then there is direction. Arabic reads right to left, which affects your entire layout, not just the text. Navigation flips. Icons that imply direction flip. Form fields, breadcrumbs and progress indicators all need to mirror. Numbers and Latin brand names stay left to right inside a right-to-left line, which is where most templates fall apart. Set dir="rtl" and lang="ar" properly at the document level and test the result on a phone, because desktop-only testing hides most of these problems.
Measuring Whether It Worked
Set this up before launch, because retrofitting it is painful and most businesses never do.
- In Search Console, filter Performance by page and use the /ar/ path to separate the two languages. Track impressions, clicks and average position independently.
- In Analytics, tag the language version so you can see enquiries by language, not just traffic by language. Traffic is not the point.
- Watch for cannibalisation in the first two months. If your English pages lost position at the moment Arabic launched, your hreflang is wrong. That is the single most useful early warning signal there is.
- Give it six months before judging. A new language section is effectively a new site section and needs time to be crawled, indexed and evaluated.
Thinking about adding Arabic to your site?
We handle the technical side, meaning URL structure, hreflang, right-to-left layout and Arabic keyword research, and we work with native speakers on the content rather than running your service pages through a plugin. Multilingual builds and SEO start at AED 1,500 per month, and we will tell you honestly if your traffic does not justify it yet.
See Multilingual SEO ServicesThe businesses that get this right treat the second language as a second website that happens to share a domain, with its own keyword research, its own content decisions and its own measurement. The ones that get it wrong treat it as a translation job, hand 140 pages to the cheapest quote, and end up with a section of their site that nobody visits and nobody has updated since launch. The technical decisions above are what separates the two, and all four of them get made in the first week.




