How Arabic-Speaking B2B Buyers Search: Intent, Dialect and Bilingual Keyword Maps

How Arabic-Speaking B2B Buyers Search: Intent, Dialect and Bilingual Keyword Maps

An English-only service page misses the queries that start a procurement conversation in Arabic. This is how those queries look, why translating your English keywords does not find them, and how to map them to pages that convert.

A supplier with a good English website and a strong reputation can be invisible to the buyer who searches in Arabic. Not because Arabic buyers do not use English; most of them do, daily. Because the query they type at the moment they need a vendor is often Arabic with the technical term left in English, and that query matches neither the English page nor an Arabic page that translated the technical term away.

This is how those queries look, why a translated keyword list does not find them, and how to build a bilingual keyword map that ends in service pages rather than a glossary.

What Arabic B2B queries actually look like

Pull a month of Search Console queries from any bilingual industrial or services site and three patterns appear.

Mixed script. An Arabic phrase with an English noun inside it: "شركة تطوير مواقع WordPress", "مورد كابلات LSZH", "تحليل SEO تقني". The buyer knows the term in English because that is how the industry uses it, and knows the surrounding words in Arabic because that is how they think. Translating the English noun into Arabic produces a query that is technically correct and rarely typed.

Intent words in Arabic, product words in English. "شركة" (company), "مورد" (supplier), "أفضل" (best), "سعر" (price), "معتمد" (certified) carry the intent. They are the words that tell you whether the query is commercial. The product term carries the topic and is often Latin.

Dialect and spelling variants. The same concept appears with different spellings and regional words: definite articles attached or not, alternative transliterations of the same English term, hamza written or dropped. A keyword tool that treats each spelling as a separate query splits the volume and hides the real size of the cluster.

Why translating your English keywords fails

  1. Word order and synonyms differ. "Backlink monitoring tool" in Arabic is not three translated words in the same order. The Arabic buyer says something closer to "tool for monitoring backlinks" with "backlinks" left in English.
  2. The volume is not proportional. A term that is huge in English may be small in Arabic and the reverse. A translated list carries the English priorities into a market with different ones.
  3. The intent markers are different words. The English list has "best", "top", "near me". The Arabic list has "شركة", "مورد", "معتمد", and location-free commercial markers like "بسعر" (at a price). Miss those and you cannot tell commercial queries from informational ones.

Build the map from Arabic sources

Work in the language, from that language's data.

  • Search Console queries, filtered to the Arabic folder, exported for 16 months. This is the only source that shows real mixed-script queries.
  • Autocomplete in Arabic, typed with the English product term inside the Arabic phrase, to see how the engine expects the query to continue.
  • Competitor titles on the Arabic pages that already rank for your category. Their H1s show the phrasing that works.
  • The sales team's inbox. How do procurement emails and WhatsApp messages phrase the request? That phrasing is the query.

Record every variant, then normalise: group spellings and dialect words under one cluster, keep the most common form as the primary, list the others as synonyms.

Cluster by intent, then assign one page each

IntentWhat the Arabic queries look likePage that owns it
Transactional"شركة" or "مورد" + product term, "سعر" + productService or product page
Commercial comparison"أفضل" + category, term + "أو" + termComparison guide, linking to the service
Informational"كيف" (how), "ما هو" (what is) + termGuide, linking up to the service
Compliance and specificationStandard number, "مواصفات", "شهادة"Specification or standards page
BrandYour name, a manufacturer you representBrand or home page

The rule that keeps the map honest: one primary URL per cluster per language. The transactional cluster belongs on the Arabic service page with the query as its H1, not on a blog post that never asks for the work. If two pages could answer it, decide which one does and link the other to it.

What the Arabic service page needs

Once the map says which query the page owns, the page has to deserve it.

  • The H1 is the query, in the buyer's phrasing, with the technical term in the script they used. Not a slogan, not a translation of the English H1.
  • Proof the reader recognises. Case studies, standards, certifications and client types that mean something to that buyer. An English-only testimonial is a weaker signal on an Arabic page.
  • A scoped next step with a number. A priced audit or a defined package converts better than a vague "request a quote" on B2B pages, in either language.
  • The technical signals. Its own URL, lang="ar", hreflang to the English twin, a self-canonical, and enough real text that the crawler does not classify it as thin.
  • The same structure as the English page, so the twin rule holds and the language switcher lands on the right place, but not the same words.

Where budgets get wasted

Writing English articles about the market and hoping the Arabic index notices. Buying an Arabic translation of the English site and calling it localisation. Building the Arabic keyword list by translating the English one and then wondering why the volumes look wrong. Each of those produces pages that exist and do not rank, and the fix is the same: research in Arabic, cluster by intent, one page per cluster, write for the query.

A worked example

An electrical systems supplier has an English page for a cable category. The Arabic queries in Search Console show "مورد كابلات" plus the cable type in English, "كابلات معتمدة" plus a standard number, and "سعر" plus the type. Three clusters: transactional, compliance, price. The map assigns the transactional and price clusters to one Arabic product-category page with the supplier phrasing as the H1 and a quotation CTA, and the compliance cluster to a standards page that lists the certifications with the products that carry them. The English twin already exists for each. The pages are linked both ways, hreflang connects them, and the sitemap lists all four. That is the entire strategy, and it fits on one sheet.

What we deliver

A keyword map per language, built from Search Console, autocomplete and competitor data, with clusters, synonyms and one primary page each. Then the content briefs for the pages that own the commercial clusters. Then, if you want, the pages themselves, written in Arabic for the Arabic query and in English for the English one.

Questions

Do B2B buyers really search in Arabic for technical products?

Yes, for a large share of commercial queries, and often with the technical term left in English inside an Arabic phrase. The query is bilingual. A page that exists only in English, or only with Arabic that translates the technical term, misses both halves.

Can I translate my English keyword list into Arabic?

You can, but the result is usually phrases nobody types. Arabic buyers use different word order, different synonyms, and English terms where the industry does. Research the Arabic queries from Arabic data sources and build a separate list that shares structure, not vocabulary.

Which Arabic dialect should a B2B website target?

Modern Standard Arabic for the page copy, because it is what every professional reads. Where dialect words appear in queries, treat them as synonyms in the keyword map and include them naturally in headings or FAQ text, not as the primary voice of the page.

Related pages

Work with Alnitak

The bilingual audit includes a keyword map per language, built from real query data, with one primary page per cluster. It is the fastest way to find out which Arabic queries you are missing.

Bilingual technical SEO auditfrom $1,700 (5–7 days)

WhatsApp this offerSee packagesContact Us

Related reading