Why Your Store Search Fails Arabic Shoppers, and How to Fix It

Arabic site search fails in most ecommerce stores because one word reaches the search box in several spellings, and a search engine that matches letters exactly treats each spelling as a different word. A shopper who types عبايه finds nothing when every product says عباية. The fix is normalisation plus a short list of synonyms built from your own zero-results report, and on Shopify it starts with knowing what the built-in search does not cover.

Normalisation is the step where a search engine rewrites both the query and the catalogue into one standard form before comparing them, so letter variants, decorative strokes and vowel marks stop mattering. English search engines have done the equivalent for decades with capital letters and plurals. Arabic needs more of it, and a store that skips it is running a search box that only works for shoppers who happen to spell exactly like its copywriter.

In Kuwait the problem is sharper than in a single-language market. The same customer may search in Arabic in the evening and in English at work, type a brand name in Arabic letters, and use a Kuwaiti word for a product your supplier described in standard Arabic. And if your store launched in English first, the Arabic catalogue was probably added later, often by a different writer, and its spellings were never checked against how people in Kuwait actually type.

Search failures are expensive because the people using the search box usually know what they want. A visitor who browses may still be deciding; a visitor who types a product name and gets an empty page has told you their intent and then been turned away. If your store gets traffic and few orders, the search box belongs on the list of exit points in why visitors don't buy, next to speed and checkout.

Letters that shoppers swap. Taa marbuta and haa are typed interchangeably at the end of a word, so عباية and عبايه, or شنطة and شنطه, are the same search. Alif maqsura and yaa get swapped too, as in كرسي and كرسى. The four alif forms, أ and إ and آ and plain ا, are often typed as plain alif on a phone, so أحذية arrives as احذية and إكسسوارات as اكسسوارات.

Hamza and its seat. Words with ؤ or ئ are often typed without the hamza, which matters to a Kuwait store selling pearl jewellery, because لؤلؤ frequently arrives as لولو. Decoration and vowel marks. Tatweel, the stretching stroke in عـطـر, and diacritics like the marks in عِطْر rarely come from shoppers. They come from product titles pasted from a designer's file, and they break matching from the catalogue side rather than the query side.

The definite article and attached letters. Arabic attaches ال, and prefixes such as و and ب and ل, directly to the word, so العطور, عطور, بالعود and للبنات each hide a product word inside a longer string. Plurals change shape as well: عطر becomes عطور, شنطة becomes شنط and حذاء becomes أحذية. A search that only matches whole strings misses all of these unless the catalogue happens to use the same form.

Digits. The Arabic keyboard on a phone can type Arabic-Indic digits, so a shopper looking for size 42 may type مقاس ٤٢ while your variants say 42. The practical test is simple: pick ten best-sellers, write each one's Arabic name in every variant above, and search your own store in Arabic on a phone, as a shopper in Kuwait would. Every empty results page is a gap your customers are already falling into.

Dialect versus standard Arabic. Suppliers and translators tend to write standard Arabic, while Kuwaiti shoppers often search the way they speak. A store that lists حذاء رياضي will miss the shopper who types جوتي, one that lists حقيبة misses شنطة, and one selling ثوب رجالي misses دشداشة. The same goes for نعال against صندل, and دخون next to بخور. A women's store selling for Gergean or Ramadan should make sure دراعة appears in its catalogue, because that is the Kuwaiti word for the garment.

English brands in Arabic letters. Kuwaiti shoppers type international brand names by sound, and there is rarely one agreed spelling. Nike can arrive as نايك or نايكي, Adidas as اديداس or أديداس, and Samsung as سامسونج, the spelling common in the Gulf, or سامسونغ, the Levantine one. If your Arabic product titles keep the brand in English letters only, an Arabic query for that brand may return nothing at all, even though the product is in stock.

Arabizi and Latin transliteration. Arabizi is Arabic written in Latin letters, with numbers standing in for sounds English has no letter for, so 3 stands for ع and 7 for ح. In a store search it shows up less as full sentences and more as product words: bukhoor, bakhoor and bakhour for incense, abaya and abaaya, dishdasha and deshdasha. A search engine configured for English treats each of those as a separate word with nothing to match.

The wrong keyboard. Some zero-results queries are not words at all. A shopper who meant to type عباية with the English keyboard still active produces ufhdm, and one who meant Nike with the Arabic keyboard active produces ىهنث. These are easy to spot in a search log once you know the pattern, and a good search tool can map them back. Where a tool can't, a handful of manual synonyms for your top products covers the queries that matter most.

Shopify's help centre page on storefront search behaviour lists the languages that get its search features, such as typo tolerance, and as of September 2026 Arabic is not on that list. The same page says typo tolerance only applies when the first four letters are typed correctly, so a three-letter word such as عطر gets no typo help even where the feature works. Shopify's developer reference for predictive search, the suggestions shown while someone types, lists its languages too, and Arabic is not among them.

That developer reference says the predictive search API returns an unsupported-locale error for languages outside its list, so check whether the dropdown on the Arabic version of your store shows suggestions at all. The full results page still works from the text in your catalogue: the help centre lists product title, product type, vendor, variant title, SKU, barcode, description and tags as the searchable product fields. Arabic wording that sits in none of those fields cannot be found.

Semantic search, which Shopify describes as using related words, concepts and categories to widen results, is included for stores with fewer than 200,000 products on the Grow, Advanced or Plus plans, and it does not apply to predictive search. The help centre notes it isn't supported for Japanese but publishes no full language list, so test Arabic queries on your own store rather than assuming they're covered. Product boosts let you attach up to ten search terms to a product so it ranks higher for them.

Synonym groups used to be the obvious fix in the Search & Discovery app, but as of September 2026 Shopify's help centre page on modifying search no longer documents them, and merchants on Shopify's developer forum reported this month that the option had gone from the app. Check your own admin before planning around it. Tags can't be translated, and the tag filter only shows in your default language, which matters if your Kuwait store is English-first. The setup in the Arabic RTL Shopify guide is the foundation here.

The most useful Arabic search research you can do is already in your admin. Shopify's analytics include a Searches with no results report, which lists the terms customers typed that returned nothing, and a Searches with no clicks report for terms whose results nobody chose. The Search & Discovery app shows the last 30 days, the full reports sit under Analytics, and Shopify notes that tracking can lag by up to 72 hours.

Export the list monthly and sort every Arabic query into five piles. Spelling variants of products you stock. Kuwaiti or Gulf words for products you stock. Brands typed in Arabic letters or Arabizi. Keyboard mistakes. And products you don't carry. The first four are search problems you can fix. The fifth is free demand research, and in a market as small as Kuwait a repeated request for something you don't stock is worth a conversation with whoever does your buying.

Fix each pile at the source where you can. Write the product's Arabic title in the form most people type, mention the Kuwaiti name in the description where it reads naturally, and add the brand in Arabic letters beside the English one. Strip tatweel and pasted diacritics from titles. Use product boosts for the few best-sellers that keep appearing. A product page that sells already needs this kind of plain, specific Arabic, so the work pays twice.

Then read the no-clicks list, because it catches a quieter failure: the search found something, just not the right thing. An Arabic query for دهن عود that returns incense burners and gift boxes ahead of the oils is a ranking problem rather than a missing word. Recheck both reports after Ramadan, Gergean and White Friday, when Kuwaiti shoppers go looking for seasonal products you may have listed under a different name.

If the native search can't cover your catalogue, a third-party search app is the usual next step, and claims on app listings are easy to make and hard to check. Ask each vendor, in writing, whether the app normalises the Arabic letter variants listed above, strips tatweel and diacritics, handles ال and attached prefixes, and treats Arabic-Indic and Western digits as the same number. A vague answer to any of these is an answer.

Then ask about the parts that matter for a bilingual Kuwait store. Can you add your own synonyms per language, including cross-script pairs such as Nike and نايك? Does the search-as-you-type dropdown work in the Arabic version of the store, laid out right to left? Does it report zero-result queries in Arabic, and can you export them? And does it support filters with Arabic labels and values rather than falling back to English?

Test before you commit. Build a list of about thirty real queries from your own zero-results report, including dialect words, brand spellings and one or two keyboard mistakes, and run every one through the trial on a phone, in Arabic. Score each result as found, wrong product or empty. That list turns a sales demo into a comparison you can repeat across two or three apps, and it takes an afternoon.

Weigh the app's cost against what it recovers, and check page speed on a phone after installing it, because a search script that slows every page can cost more than the search wins back. If your store also competes in Google, the same spelling list feeds your content work in Arabic SEO. If you're building the store now, you can start a free Shopify trial and test Arabic search on your own catalogue before launch.