The most expensive mistake I see in this region is running English keyword research, translating the output, and calling it an Arabic strategy.
It fails for a specific reason. Keyword translation preserves the words and loses the intent. The English term with the volume frequently maps to an Arabic phrase nobody uses, while the phrase people actually type, often in dialect and often mixing Arabic and Latin script, never appears in the research at all.
What makes Arabic search different
Dialect versus Modern Standard Arabic
Egyptian, Gulf and Levantine users search differently. MSA content can rank and still convert badly because it does not sound like the reader.
Script and transliteration
Users mix Arabic and Latin script, drop diacritics inconsistently, and often search brand names transliterated. Your keyword set has to account for all of it.
Thin SERP competition, high technical debt
Many Arabic SERPs are less competitive than their English equivalents. The opportunity is real, but often the blocker is your own site’s RTL and hreflang handling.
Intent splits by market
The same Arabic phrase can carry commercial intent in Riyadh and informational intent in Cairo. One page for both usually serves neither well.
What the work involves
- Native Arabic keyword research, built from search data rather than translated
- Dialect and market segmentation across Egypt, Saudi Arabia and the UAE
- Intent mapping per market, not per keyword
- hreflang architecture and validation
- URL structure decisions for bilingual sites
- RTL rendering and layout auditing
- Content parity strategy: what needs translating, what needs writing fresh
- Arabic content briefs your writers can work from
- Transliteration and brand-name variant coverage
- Local entity and Google Business Profile work in Arabic
The bilingual architecture decision
There are three workable structures for an Arabic and English site, and the wrong one is expensive to undo later:
Subdirectories
/ar/ on the same domain. Strongest option for most businesses: authority is shared, maintenance is simplest, hreflang is straightforward.
Subdomains
ar.example.com. Justifiable when the two sites are genuinely separate operations with separate teams. Splits authority.
Separate domains
Rarely the right answer unless there is a legal or brand reason. Doubles the work and divides everything.
For a consulting or service business serving all three markets, subdirectories are almost always correct. If someone has already built you the other way, migrating is a real project, worth doing, but plan it properly.
Frequently asked
Should we translate our English content or write new Arabic content?
Neither, exclusively. Reference and product pages usually translate fine. Anything targeting a search query should be researched and written from Arabic keyword data, because the topic that has demand in Arabic is often not the topic that has demand in English.
Do we need separate content for Saudi Arabia, Egypt and the UAE?
For most informational content, no. For commercial pages where pricing, regulation, delivery or local terminology differ, yes. Hreflang with regional targeting is how you do it without creating duplication problems.
Our Arabic pages are not indexed. Why?
Most often one of three things: hreflang pointing at the wrong URLs, Arabic pages canonicalising to their English equivalents, or the Arabic version being generated client-side so the crawler never sees the content. All three are findable in a couple of hours.
Can you write the Arabic content?
I write the briefs, the keyword mapping and the on-page structure, and I review the final copy. For volume writing I work with native writers per dialect, which produces better copy than any single person covering three markets.