You spent weeks building the perfect Arabic page, translated every keyword flawlessly, and ranked for terms nobody types. Somewhere in Riyadh, a buyer with cash in hand just searched عروض الجوالات (mobile offers) and landed on your competitor — because they wrote for how people actually speak, and you wrote for a grammar textbook. In Arabic search, the gap between the right word and the almost-right word isn’t a few rankings. It can be the difference between 50 searches a month and 50,000.
I’ve watched this play out across Egypt, Saudi Arabia and the Gulf for years. The single most expensive mistake I see — made by smart marketers, big agencies, and confident founders alike — is treating “Arabic” as one keyword market. It is not. This guide is the practical, field-tested process I use to find the Arabic keywords real buyers type, market by market, dialect by dialect. No theory, no fluff — just the workflow that took a Saudi store from roughly 25 impressions a day to 51.5M impressions, and rebuilt a collapsed niche store to #1 in Saudi Arabia in 166 days.
Arabic is not one search language
Here is the foundational truth almost every Arabic SEO failure traces back to: there is no single “Arabic” your customers search in. What we call Arabic is really a family — at least five major regional varieties plus a formal written standard. Each one shapes how people phrase a query into the search box.
The five major dialect groups, plus Modern Standard Arabic, are:
- Gulf (Khaleeji) — Saudi Arabia, UAE, Kuwait, Qatar, Bahrain, Oman
- Egyptian (Masri) — Egypt, and widely understood across the region thanks to film and TV
- Levantine (Shami) — Syria, Lebanon, Jordan, Palestine
- Maghrebi (Darija) — Morocco, Algeria, Tunisia, often mixed with French
- Mesopotamian (Iraqi) — Iraq
- Modern Standard Arabic (MSA, or Fusha) — the formal written register used in news, official content and pan-regional writing
The trap is that almost all formal Arabic education, and almost all machine translation, defaults to MSA. So when a marketer “translates the keywords into Arabic,” they produce textbook Fusha. The problem? People don’t type Fusha into Google for everyday shopping and questions. They type the words they actually speak. A Saudi looking for a phone deal doesn’t search the grammatically pristine MSA phrase for “best smartphone offers” — he searches عروض الجوالات, plain Gulf-dialect everyday speech. Optimize only for MSA and you’ve optimized for how almost nobody actually searches.
This doesn’t mean MSA is useless — far from it. For news, formal informational content, and pan-regional reach, MSA is exactly right. A single well-written Fusha article can earn visibility from Morocco to Iraq. The skill is knowing when to reach for Fusha and when to reach for dialect — and that decision is driven by intent, which we’ll come to. For now, internalise the first rule: research Arabic as a set of markets, never as one.
Why dialect choice changes everything
If dialect were just a stylistic preference, you could shrug it off. It isn’t. Dialect choice can change search volume by orders of magnitude — and that is the whole ballgame for keyword research.
The textbook example circulates across the MENA SEO industry for a reason: the same shopping intent can return roughly 50 monthly searches in one phrasing and around 50,000 in the regional dialect phrasing. Gulf shoppers hunting for phone deals overwhelmingly type عروض الجوالات (literally “mobile offers” in everyday Gulf speech) rather than the literal MSA translation of “best smartphone deals.” Target the Fusha phrasing and you’re competing for crumbs while your competitor feasts on the dialect term — for the exact same buyer.
This is why I never accept a keyword list at face value. The “obvious” translation is frequently the low-volume one. Watch how a single product can fracture across markets:
| Intent | Modern Standard Arabic (Fusha) | Common dialect / market term | Where volume lives |
|---|---|---|---|
| Mobile phone offers | أفضل عروض الهواتف الذكية | عروض الجوالات | Gulf (جوال is the Gulf word for phone) |
| Mobile phone offers | أفضل عروض الهواتف الذكية | عروض الموبايلات | Egypt (موبايل is the Egyptian word) |
| Car for sale | سيارة للبيع | عربية للبيع | Egypt (عربية = car colloquially) |
| Apartment for rent | شقة للإيجار | شقق للايجار (undiacritized, plural) | Both, but spelling varies |
| Laptop | حاسوب محمول | لابتوب | Everywhere — the loanword wins |
Notice three things. First, the Gulf word for “phone” (جوال) and the Egyptian word (موبايل) are completely different — a single Arabic page targeting one will miss the other’s market entirely. Second, English loanwords like لابتوب (laptop) often outperform the “correct” Arabic term, because that’s what people actually say. Third, the formal term is almost never the high-volume one for transactional intent.
SEO localization goes beyond simply translating content — it involves tailoring it to your target audience’s local culture, preferences, and search habits.
This is exactly the principle behind the niche store I rebuilt to #1 in Saudi Arabia in 166 days. We didn’t just translate the old keyword list — we threw it out and rebuilt it around the Gulf-dialect terms Saudi buyers actually typed, market-specific spelling variants and all. The rankings followed the language, not the grammar.
Never translate keywords word-for-word
I want to be blunt here because it costs people so much money: do not translate your English keywords into Arabic with a tool and call it research. Word-for-word machine translation produces phrases that are grammatically correct and behaviourally wrong. They match the dictionary, not the searcher. Worse, they can map your content to the entirely wrong intent — you end up ranking, if at all, for a query no buyer means.
There’s a second, harder consequence. Google’s spam policies explicitly target content that’s been machine-translated and dumped onto pages without human review or added value.
Scraping feeds, search results, or other content to generate many pages (including through automated transformations like synonymizing, translating, or other obfuscation techniques), where little value is provided to users.
So auto-translation isn’t just ineffective — at scale, it’s a risk to your whole site. The recurring expert consensus, and my own practice, is the same: research seeds with a native speaker for each target market. A native speaker doesn’t translate; they tell you what people actually say. They know that in Egypt a car is عربية in casual speech but سيارة in a formal listing. They know which spelling of شقق للايجار people type without diacritics. They know the slang, the loanwords, and the code-switching — the mixing of Arabic and English, sometimes written in Latin letters (Arabizi), that mainstream tools simply don’t capture.
Translation has its place — for adapting content once the strategy is set, always with human editing. But the keyword strategy itself must be built from how people search, not from a bilingual dictionary. This is the difference between localization and translation, and it’s the line that separates pages that rank from pages that don’t.
The tools — and their Arabic blind spots
You still need tools. They give you volume, difficulty and competitive data no human can eyeball. But you have to use them knowing that most mainstream SEO tools were built for English and handle Arabic imperfectly. They frequently ignore diacritics, conflate dialects into one undifferentiated “Arabic,” and miss code-switched or Arabizi queries entirely. Treat their numbers as directional, not gospel — and always finish with manual validation.
Here’s the stack I rely on, and exactly what each is good and bad at for Arabic:
| Tool | Best for in Arabic | The blind spot to watch |
|---|---|---|
| Google Keyword Planner | Geo-specific volume when you set location + language explicitly (e.g. Saudi Arabia + Arabic) | Bands volume into wide ranges; mixes dialects unless you feed dialect seeds |
| Ahrefs | Large keyword database, difficulty, competitor gaps | Can conflate spelling variants; verify dialect terms manually |
| Semrush | Broad database, intent labels, competitor analysis | Arabic data thinner than English; cross-check volume |
| AnswerThePublic | Arabic question phrasing for long-tail and AI clusters | Coverage varies by market; best as an idea generator |
| Google Trends | Regional comparisons (KSA vs Egypt vs UAE) over time | Relative, not absolute volume — use for direction |
| Google Autocomplete | Real queries on country domains (google.com.sa, google.com.eg) | Manual; but the most honest signal of how people actually type |
The non-negotiable step everyone skips is the last one: manual validation. Open google.com.sa and start typing your seed in Gulf dialect — Autocomplete shows you the real long-tail people search in Saudi Arabia. Repeat on google.com.eg for Egypt and watch how the suggestions change. Then test multiple orthographies: alef and hamza variants (إ vs ا vs أ), the undiacritized spelling versus the textbook one, and stop words attached to the next word the way Arabic actually writes them. The “same” keyword can be written several ways, and the high-volume version is usually the casual, undiacritized one — not the spelling your tool defaults to.
For Keyword Planner specifically: set the location and the language as two separate, explicit settings. “Saudi Arabia + Arabic” returns a very different landscape from “Egypt + Arabic.” Feed it dialect seeds, not English-to-Arabic translations, and you’ll see volume the default workflow hides.
Build per-market keyword maps, not one Arabic list
By now the conclusion writes itself: don’t build one Arabic keyword list. Build a keyword map per market. Saudi Arabia, Egypt and the UAE are not three accents of one audience — they are three distinct intent landscapes with different dialects, different spellings, different brand names, different common product terms, and wildly different competition and reach.
The reach gap alone is dramatic. Saudi Arabia entered 2025 with near-universal internet penetration — around 99%, roughly 33.9 million users — making it a fully-online, high-competition search market where ranking is hard but every position is valuable. Egypt has the largest social audience in MENA at 50.7 million users, yet penetration trails the Gulf and masks a wide urban-rural connectivity gap. The same Arabic keyword can therefore have very different reach, competition and conversion potential depending on which market you’re in.
So my per-market map for any project looks like this:
| Market | Primary dialect | Volume / competition profile | Practical note |
|---|---|---|---|
| Saudi Arabia (KSA) | Gulf (Khaleeji) | High penetration (~99%), high competition | Premium buyers; dialect terms like جوال; mobile-first |
| Egypt | Egyptian (Masri) | Huge audience, uneven connectivity | Price-sensitive; terms like موبايل, عربية; big long-tail |
| UAE | Gulf, but very multilingual | High value, often bilingual searches | Heavy English/Arabic code-switching; test both |
Each row gets its own seed list, its own Autocomplete validation, and its own intent classification. Sometimes the markets share a head term and diverge only on the long tail; sometimes the core product word itself differs. You only learn which by researching each market on its own terms.
This is precisely the structure I used on Conscent, which grew from 61K to 1.2M impressions once the keyword strategy was rebuilt around the right market and the right phrasing — and on Oxford in Egypt, where Egyptian-Arabic targeting did the heavy lifting that a generic Arabic list never could. If you sell across borders, your e-commerce SEO has to respect those borders inside the keyword map, or you’ll bleed traffic at every one.
Intent comes first, always
Volume is seductive. But before search volume, before difficulty, before you build a content calendar, one question outranks all the others: does the intent behind this keyword match something my site can actually serve and convert?
Before you look at keyword search volume, assess keyword difficulty, or plan a content calendar, ask whether the intent behind a keyword matches something your site can realistically serve to searchers and convert.
In Arabic this matters even more, because dialect and formality often signal intent. Classify every keyword into the four classic buckets:
- Informational — “how do I…”, “what is…”. Often phrased as a natural-language question, frequently in MSA for pan-regional reach. Serve with guides and answers.
- Commercial — “best…”, “review…”, “X vs Y”. Buyers comparing options, usually in dialect. Serve with comparisons and category pages.
- Transactional — “buy…”, “offers…”, “price…”. Ready to act — and almost always in dialect (عروض, سعر, شراء). This is where money lives.
- Navigational — searching for a specific brand or site. Capture with branded pages.
Notice the pattern: the closer to a purchase, the more dialect-heavy the query. Nobody types a formal Fusha phrase when they’re about to buy a phone — they type عروض الجوالات. So your transactional research must be dialect-first, while your informational research can lean on MSA for breadth.
And then there’s the structural reality of Arabic search: the overwhelming majority of it is long-tail. This isn’t an Arabic quirk — it’s how search works everywhere, but it’s amplified by Arabic’s conversational, question-rich phrasing.
Long tail searches aren’t just longer. They’re also more specific. They usually have commercial or transactional search intent.
With around 92% of all keywords earning ten or fewer searches a month, the bulk of demand isn’t in a few fat head terms — it’s spread across thousands of specific, intent-rich long-tail phrases. In Arabic, that means question-shaped and conversational queries win. Chase the giant head term everyone fights over and you’ll lose to bigger sites; build out the long-tail, dialect-specific cluster around real buyer questions and you’ll quietly own the traffic that converts.
AI search and voice are now Arabic-first concerns
If you think AI search is an English-only story, you’re already behind. Google rolled out AI Overviews to 40+ languages — including Arabic — in May 2025, and launched AI Mode in Modern Standard Arabic in October 2025. Arabic SERPs now routinely carry AI-summarised answers above the classic results, and that changes what keyword research has to capture.
AI Overviews are now available in over 200 countries and territories, and more than 40 languages.
The behavioural shift is the part to internalise: Arabic-market AI queries are reported at two to three times the length of traditional searches. People aren’t typing two-word fragments to an AI — they’re asking full, conversational questions, often with follow-ups. That makes question-shaped, long-tail keyword clusters far more valuable than short head terms, and it rewards research that maps natural-language questions and their follow-up intents.
Voice search pushes in the same direction — and it accelerates the move from MSA to colloquial Arabic. When someone speaks to their phone, they don’t speak Fusha; they speak their daily dialect, in full natural sentences. So voice and AI both reward the same thing: natural-spoken, long-tail, question-based phrasing that varies by dialect. Your research should explicitly harvest these — open-ended “كيف”, “ليش”, “إزاي”, “أفضل” questions in the right dialect, mapped into clusters an AI answer can pull from. This is where AnswerThePublic and Autocomplete earn their keep, surfacing the real conversational phrasing people use.
The strategic point: chasing AI visibility in Arabic doesn’t require some secret new technique. It requires the same intent-first, dialect-aware, long-tail research — just weighted harder toward conversational, question-shaped clusters than ever before.
The technical foundation behind every keyword
Keyword research only pays off if Google can correctly read your page and match it to the right local query. That’s a technical SEO job, and in Arabic it has a few non-negotiables.
Start with how Google identifies a page’s language:
Google uses the visible content of your page to determine its language. We don’t use any code-level language information such as lang attributes, or the URL.
Read that carefully, because it has a sharp consequence: the visible Arabic copy on the page does the heavy lifting. Machine-translated, awkward, or wrong-dialect text will be read as exactly that — and won’t match the local queries you researched. Natively written Arabic in the right register is the technical requirement, not a nice-to-have.
The rest of the foundation:
- Right-to-left (RTL) layout, done properly so the page renders and reads naturally for Arabic users — not a left-to-right template with text crammed in backwards.
- Separate URLs per language version, so each language has its own indexable address.
hreflangannotations —arfor general Arabic, plus regional tags likear-sa,ar-egorar-aewhere the intent and phrasing genuinely diverge between markets. This tells Google to serve the Saudi page to Saudi searchers and the Egyptian page to Egyptian ones.- Natively written content, not machine translation — both for ranking and to stay clear of Google’s spam policies.
Get this layer right and your dialect-specific keyword work actually connects to dialect-specific traffic. Get it wrong and even perfect research is wasted, because Google never matches your page to the searcher you wrote it for. The technical layer is where Roseberry’s climb to 51.5M impressions truly began — a 17-point technical fix was the foundation that let the Arabic keyword and content work compound.
My step-by-step Arabic keyword research workflow
Here’s the whole process, in the order I run it, so you can apply it yourself:
- Define the markets. Decide exactly which countries you’re targeting — KSA, Egypt, UAE, and so on. Each becomes its own research track, never a shared “Arabic” bucket.
- Run a native-speaker seed session. For each market, brainstorm with a native speaker how a real customer describes the product out loud — slang, loanwords, dialect, spelling shortcuts. These raw human seeds are the foundation; tools only refine them.
- Validate seeds in Google Autocomplete on the country domain (
google.com.sa,google.com.eg). Capture the real long-tail and note how phrasing shifts between markets. - Test orthography variants. For each promising term, check alef/hamza variants and the undiacritized spelling. Keep the version people actually type.
- Pull volume in Keyword Planner with location + language set explicitly per market, fed with dialect seeds — never English translations.
- Expand and compete with Ahrefs / Semrush. Find related terms, difficulty and competitor gaps, treating tool numbers as directional and verifying dialect terms by hand.
- Harvest question clusters with AnswerThePublic and Autocomplete for AI and voice — natural-language, conversational, dialect-correct questions and follow-ups.
- Classify by intent — informational, commercial, transactional, navigational — keeping transactional research dialect-first.
- Build the per-market keyword map, matching each keyword to a specific page type before any content is written.
- Wire up the technical layer — RTL, separate URLs, regional hreflang, natively written copy — so Google matches each page to the market you researched it for.
Follow that and you stop guessing. You stop ranking for textbook phrases nobody types. You start showing up exactly where the buyer with cash in hand is already searching — in his dialect, on his country’s Google, for the thing he’s about to buy.
Arabic keyword research isn’t harder than English keyword research — it’s just less forgiving of shortcuts. The marketers who win MENA search aren’t the ones with the biggest tool subscriptions. They’re the ones who respect that Arabic is a family of markets, who research how people actually speak, and who build the technical foundation that connects the two. That’s the work behind every result I’ve shared in this guide — and it’s exactly the work I can do for your market.
Frequently asked questions
If you take one thing from this guide, let it be this: in Arabic search, the right word and the almost-right word are not close. One reaches the buyer; the other reaches a grammar textbook. Choose deliberately, market by market — and if you’d rather have it done right the first time, my SEO services are built for exactly the Egyptian, Saudi and Gulf markets this guide describes.