Name the garment the way she types it, not the way your tag spells it
Start with the word, not the tag. A woman in Kuwait searching for an abaya types عبايات, not abayas. She types دراعات for the everyday dress your site calls a kaftan, فستان سهرة for what your line sheet calls an evening gown, and جلابية for what the photographer labelled a house dress. Your product titles were written off the supplier invoice, usually in English, usually in fashion-trade vocabulary. Nobody searches in fashion-trade vocabulary. Pull your last thirty WhatsApp and Instagram DM enquiries and write down the exact Arabic nouns the customers used. That list is your keyword file, not a tool's export.
Then handle the mixed spellings, because Kuwait types both ways. Plenty of shoppers write دريس, بدي سوت, كروب توب, English garment words spelled in Arabic letters, and just as many write the brand name in English inside an Arabic sentence. Both belong on the page. Put the Arabic noun in the H1 and the title tag, and let the transliteration and the English word sit naturally in the body copy and the product specs. That is not keyword stuffing; it is covering the three ways one garment gets named. The searches for a single item split across all three, and most sites capture one.
Ready to start your Shopify store?
Start a free trial and try the platform for yourself.
Eid and wedding season are search seasons, and they start six weeks early
Occasion queries carry the money in this category. ملابس عيد climbs roughly three to four weeks before Eid al-Fitr and holds until the night before. فساتين سهرة runs all year but doubles through wedding season and again around Hala February. عبايات رمضان and دشاديش رمضان start moving the moment Sha'ban does. And ملابس عيد للبنات is a separate query from ملابس عيد that converts better, because mothers buy earlier and buy in twos. Build one page per occasion, not a blog post. A page with products on it, live all year, restocked and refreshed each season.
Publish those pages six weeks before the traffic, not the week of. Google needs time to index and rank a new URL, and an Eid page posted three days before Eid earns nothing. Keep the same URL every year and swap the products inside it; a page living through its third Ramadan outranks a fresh one every time. Expect the occasion pages to carry 30 to 60 percent of your organic sessions in the four weeks around Eid, and plan stock and Snapchat budget against that. Then in July and August expect the whole thing to go quiet while Kuwait travels.
A translate widget is not an Arabic site
Arabic needs its own URLs, /ar/abayat, not a JavaScript toggle that swaps the text on the same address. If the Arabic version has no URL of its own, Google has nothing to index and you rank for nothing in Arabic. Set hreflang between the English and Arabic versions, use ar-KW on the Kuwait pages, and make sure the Arabic H1, title tag and meta description are actually Arabic and not the English ones left in place. Then add 150 to 250 words of real Arabic copy on every category page, because a grid of images with no text has nothing to rank on.
Product descriptions have to be written, not machine-translated. Auto-translation mangles قماش كريب, gets sizing wrong, and reads like a robot to a woman deciding whether to spend 45 KD. Write four Arabic sentences per product: fabric, fit, length in centimetres, and what she would wear it to. Budget 1 to 2 KD per product for a native writer, or write them yourself in batches of twenty. If your platform cannot give Arabic its own indexable pages, change platforms, because Shopify runs Arabic as a real second storefront with separate URLs, its own meta fields, and KNET at checkout.
Here is the opening. Walk through ten Kuwaiti fashion sites and eight will be English-only at product level, maybe with an Arabic homepage bolted on top. Meanwhile most fashion searching in Kuwait happens in Arabic, on a phone. That gap is why a small brand with 200 properly written Arabic product pages can outrank a much bigger name within six to nine months. It is slow, boring work, roughly 20 to 40 hours to convert a 200-product catalogue, and that is exactly why almost nobody in the market has done it.