The short answer: the three tiers, in order
Localise in this order. Tier 1: your money pages — homepage, core service pages, pricing and contact — plus any page that already earns significant traffic or enquiries in English. Tier 2: high-intent guides and location pages where Arabic searchers are actively comparing providers. Tier 3: support content — company background, careers, low-traffic posts — and anything you should deliberately defer, including pages with no Arabic search intent at all. Score every page first using the model below, phase the work by tier, and decide page by page whether the Arabic version should be a translation, a rewrite, or written Arabic-first. The first ten Arabic pages should be the ten most likely to win enquiries, not the ten easiest to translate.
The problem: localisation budgets spent in the wrong order
The usual pattern is backwards. The translator works through the site in menu order, or starts with the shortest copy, and the budget runs out just before the service pages that matter. Meanwhile the English site keeps growing, so the Arabic version falls behind and often covers only a minority of pages — with the commercial core the part most likely missing. The Arabic section looks complete in the navigation but cannot do the job it was funded for: turning Arabic searchers into enquiries. Sequencing is the fix, and it costs nothing except deciding first.
The scoring model: Arabic intent × commercial value × current performance
Score each page from 1 to 5 on three factors. First, Arabic intent: is there evidence people search in Arabic for what this page offers — Arabic queries in Search Console, Arabic keyword research, or enquiries already arriving in Arabic? Our Arabic keyword research guide covers gathering that evidence. Second, commercial value: how close is the page to revenue? A service page scores 5; an awareness post scores 1 or 2. Third, current performance: does the English page already attract traffic or conversions worth extending into Arabic?
Add the scores. Pages at 12–15 are Tier 1 candidates; 8–11 form Tier 2; 7 or below wait. The model is deliberately simple — it forces the evidence conversation before money is spent, and makes deferral a decision rather than an accident.
Tier 1 — money pages and top landing pages (with the evidence to pull)
Tier 1 is short by design: homepage, the service pages carrying your offers, pricing, and contact or booking pages — plus any English page already earning significant traffic. These are the pages Arabic visitors need to complete a journey, and where an Arabic version can rank for commercial queries. Our own Arabic SEO service page at /ar/seo-services-uae exists because that intent is commercial, while the English SEO services page serves the English journey.
Before commissioning Tier 1, pull per page: organic landing traffic and conversions, any Arabic queries already reaching it, and its role in assisted conversions. If a money page has no traffic and no Arabic intent, fix the page first — translating a weak page produces a weak Arabic page.
Tier 2 — high-intent guides and location pages
Tier 2 holds content that supports buying decisions: comparison guides, cost guides, checklists and location pages for the emirates or districts you genuinely serve. These pages win Arabic informational searches from people who are weeks, not months, from enquiring. The bilingual structure principles in our guide to SEO for bilingual Arabic-English websites apply here, but the sequencing point is simpler: localise the guides that sit closest to your Tier 1 services first, so Arabic readers can move from guide to service page without switching language mid-journey. A brilliant Arabic guide that links only to English service pages leaks most of its value at the last click.
Tier 3 — support content, and what to deliberately defer
Tier 3 is everything else: company history, team pages, news, low-traffic blog posts and legal detail beyond what customers need to transact. Some of it should be localised eventually; much of it can wait a year without costing an enquiry. Defer deliberately, and write the defer list down. Pages with no Arabic intent, pages about to be rewritten in English, and pages kept only for completeness belong on it. So do near-duplicate location pages — localising thin content multiplies the problem in a second language. A smaller, complete-feeling Arabic site beats a sprawling half-finished one, for visitors and for search engines alike.
Translate, rewrite, or write Arabic-first? Deciding per page
Not every page should be handled the same way. Translate directly when the English page is factual and stable — contact details, process steps, specifications. Rewrite when intent differs in Arabic: the questions Arabic searchers ask, the terminology they use, and the objections they raise often are not the ones the English page was built around. Write Arabic-first when the offer itself is Arabic-led; a service aimed at Arabic-speaking customers deserves an original Arabic page, with English as the adaptation if one is needed.
Use the scoring evidence: high Arabic intent with different query patterns means rewrite or Arabic-first; matching patterns mean a careful translation is enough. Either way, the Arabic page must stand complete in its own right — which is why implementation details such as layout direction matter, as our Arabic RTL website checklist sets out.
Worked example: a 60-page UAE services site, phased
Illustrative example — a hypothetical site, shown to demonstrate the method, not a real client or measured results.
Imagine a UAE services firm with 60 English pages: 8 service pages, 6 location pages, 30 blog posts, and 16 support pages. Scoring puts 11 pages in Tier 1 — homepage, 6 service pages with proven Arabic intent, pricing, contact, and 2 top landing posts. Phase one localises those 11: 7 rewritten around Arabic intent, 4 translated. Phase two takes 14 Tier 2 pages — location pages and the 8 guides closest to Tier 1 services. The other 35 are deferred as a written decision, reviewed quarterly. Initial scope: 25 of 60 pages — but they include every page an Arabic customer needs to find, evaluate and contact the firm.
Rules that protect the English site while you localise (equivalence, internal links, maintenance)
Three rules keep localisation from damaging what works. First, equivalence: pair English and Arabic pages as versions of each other only when they genuinely cover the same ground; a rewritten Arabic page with different scope is a separate page, not a twin. Second, internal links: Arabic pages should link to Arabic pages wherever the destination exists, so visitors are not bounced into English mid-journey. Third, maintenance: every localised page doubles its content's update cost, so agree who updates the Arabic version when prices or services change. Stale Arabic pages undo the trust localisation was meant to build. Newer search surfaces reward the same discipline — consistent, well-maintained pages are what our AI visibility work builds on.
Common mistakes
Translating in menu order and exhausting the budget before the service pages. Translating everything literally, including pages whose Arabic intent differs. Commissioning original Arabic content for pages nobody searches for in Arabic. Arabic landing pages that link only to English contact forms. And launch-and-forget: an Arabic site never updated, slowly contradicting the English one. Every one is a sequencing or governance failure, not a translation-quality failure.
FAQ
How many pages should we localise at launch?
Enough to complete the customer journey in Arabic — typically your Tier 1 set: homepage, core services, pricing and contact, plus top landing pages. For many UAE SME sites that is 8–15 pages, not the whole site.
Should we translate our blog into Arabic?
Selectively. Localise only posts with demonstrated Arabic intent or a close link to a Tier 1 service — usually a minority of the blog. Translating a whole archive in order is the most expensive way to gain the least visibility.
Is machine translation good enough for Arabic pages?
For drafts or low-value support content, possibly with human review. For money pages, raw machine translation usually misses Gulf business terminology and reads as translated, undermining trust. Tier 1 deserves professional translation, rewriting, or Arabic-first writing.
What if a page has no Arabic search demand?
Defer it, and record the decision. No Arabic intent means no Arabic search visibility, however good the translation. Revisit quarterly — demand changes, and your Arabic presence can reveal intent the data did not show at first.
Do Arabic pages need separate keyword research?
Yes. Arabic queries differ in wording, terminology and intent, so each page needs its own Arabic keyword check first — translating the English keywords produces pages optimised for searches nobody makes.
Get an Arabic localisation roadmap (CTA)
If your Arabic version is planned as one big translation job, pause and sequence it first. WebStackRank's SEO services for the UAE include the prioritisation behind this framework: scoring your pages, then phasing localisation so the first dirhams land on the pages most likely to win enquiries. Ask for a roadmap before commissioning translations.
