E-commerce development · Gulf
Online Stores Built to Sell in the Gulf
Arabic and English stores with Mada, Apple Pay, instalments and cash on delivery configured and tested before launch. Fixed quotes, and a checkout built for how the Gulf pays.
AED 16,500
online store, from
Bilingual catalogue, local payment methods configured, hosting for the first year
6–8 wks
to launch
A focused catalogue. Product data sets the pace, not development
4 markets
four sets of rails
The UAE, Saudi Arabia, Qatar and Kuwait each need their own set-up
Arabic
product data, not labels
Names, attributes, filters, invoices and notification templates, not only descriptions
What changes when you sell in the Gulf
Key takeaways
- Cash on delivery is not a payment method. It is unsecured credit, and the refused delivery is the bill.
- A Saudi store without Mada declines the country’s most widely held card.
- Apple Pay converts because it deletes the two worst steps of a mobile checkout: the address form and the card form.
- WhatsApp is where the sale happens. A chat button nobody answers is worse than no button.
- A required postcode field is a bug in the UAE and Qatar. A map pin beats any line of the address.
What is on this page
A Gulf store is not a Western store translated
A store built for London carries assumptions in its bones. The buyer has a card saved. The address has a postcode that resolves to a street. A courier leaves the parcel at the door.
Every one of those is wrong here, or half true, and translating the interface touches none of them. You can have a flawless right-to-left layout and still lose the order at the payment step, because the card in the buyer’s hand is a Mada card and your checkout has never heard of it.
So the work is mechanical, not cosmetic. Six things behave differently: how people pay, what happens when they pay cash at the door, which channel the conversation runs on, what Arabic product data means, how an address and a phone number are written, and what a buyer expects when they send something back. For a brochure site instead, see web design; for traffic to a store you already have, SEO and Google Ads.
Payment and checkout options by market
What buyers expect, what you cannot launch without, and what it costs you. We do not quote rates: your acquirer and volume set them, so get two written offers.
| Market | What buyers expect | Effectively mandatory | What it costs you |
|---|---|---|---|
| United Arab Emirates | International cards, Apple Pay, cash on delivery, instalments split into four | Apple Pay and cash on delivery | A gateway normally needs a UAE entity and bank account. Cash orders add a courier fee, plus the return leg on refusal. Instalments carry a merchant fee. 5% VAT on prices and invoices |
| Saudi Arabia | Mada debit first, then Apple Pay, wallets such as STC Pay, instalments, and cash on delivery outside the main cities | Mada. Without it you decline the commonest card in the country | Mada acquiring through a Saudi gateway, and a checkout that routes Mada card ranges and survives the one-time code. 15% VAT, and invoices meeting the national e-invoicing rules |
| Qatar | Domestic debit rails, international cards, Apple Pay, cash on delivery | Domestic debit and cash on delivery | Gateway access is arranged through a Qatari bank. No VAT to display at present. A smaller courier network means longer delivery windows, which raises cancellations on cash orders |
| Kuwait | KNET before anything else, then cards and Apple Pay, with cash on delivery common | KNET. A card-only checkout is a non-starter | KNET is integrated through a Kuwaiti bank, priced per transaction rather than as a percentage. No VAT. Addresses arrive as block, street and building, so the form is rebuilt |
You cannot buy your way past the entity question: local acquiring usually needs a licensed presence, which is your accountant’s decision. And every method you add is one somebody reconciles monthly, so three that work beat six that half-work. Prices sit on the Dubai cost page and the Saudi one.
Cash on delivery, and what a refusal costs
Every platform lists cash on delivery as a payment method. It is not one. It is an unsecured, interest-free loan to someone you have never met, repayable at their door, at a time of their choosing, cancellable without penalty. Priced properly it is often still worth offering. Priced like a card payment, it removes the margin from a whole category.
A refusal is not one cost. You pay the outbound leg, the return leg and handling at both ends, and the stock sits in a van instead of on a shelf for days, which hurts most on your fastest lines.
The reasons are boring and mostly fixable: a mistyped number, an address the courier cannot find, nobody home in the window, a duplicate order, an impulse that faded. We will not tell you what share of your orders will be refused; anyone quoting a figure for your catalogue and your courier is guessing. What reduces refusals is concrete: confirm before packing, validate the phone at checkout, cap the value payable in cash, charge a handling fee so prepayment is cheaper, let the buyer pick the window, and keep a rule for repeat refusers.
The confirmation step, drawn
The same order twice: once dispatched on trust, once confirmed before anything is packed.
The confirmation is unglamorous: one message in Arabic and English naming the item, the total and the delivery day, with one tap to confirm and one to cancel. Send it on WhatsApp: it will be read. Hold the order pending until it is confirmed, and make cancelling free so people use it rather than ignore you.
Mada, Apple Pay and instalments
Mada is Saudi Arabia’s domestic debit scheme and the card most Saudi buyers hold. Many Mada cards are co-badged with an international scheme and work online through that route; plenty are not, and those fail on a checkout that only knows Visa and Mastercard. Supporting it means acquiring through a Saudi gateway, recognising Mada card ranges, and not dropping the basket when the one-time code is late.
Apple Pay is the highest-converting method we configure, for a structural reason rather than a cultural one: it removes the address form and the card form, the two steps where mobile checkouts bleed.
Instalments are what Western-built stores miss. A provider splits the total into four payments, pays you in full, and charges the merchant a fee per order higher than card processing. The buyer pays nothing extra. That fee is the whole trade: a larger basket on considered purchases, nothing at all on cheap ones. It has to show on the product page with the per-instalment amount, in Arabic, not only at checkout. Salla and Zid arrive with all of this wired in.
WhatsApp is a sales channel, not a support widget
Most stores bolt a chat bubble to the corner, route it to an inbox and call it support. That is a misuse of the most important channel you have. A large share of buying conversations here start, continue and finish on WhatsApp.
Wiring it in is a handful of decisions. Put click-to-chat on the product page, not only the footer, and pre-fill the message with the product name and code so nobody has to describe what they are looking at. Send order, dispatch and cash-confirmation messages into the thread the buyer already opened.
Then the hard part, which is not technical. Someone has to answer, in Arabic, within minutes, in Gulf hours. An unanswered button does more damage than none, because it turns an interested buyer into an ignored one. One handset works until you have two people; past that you need the business API, several agents on one number and a record of what was promised. The same logic runs through our interface design and campaign work.
Where a mobile checkout loses people
A default checkout, as most platforms ship it, with the four steps that cost Gulf stores orders.
None of these are design problems. They are configuration and validation problems: cheap to fix, and verifiable on a real phone. Test on a mid-range Android handset on mobile data, and place one real order per payment method before launch, then refund it. Google’s Core Web Vitals guidance covers the speed half of the same problem.
Arabic product data, not an Arabic skin
Most bilingual stores translate descriptions and stop. The buyer then meets a catalogue where the product name is Arabic, the colour is “Navy”, the filter is labelled “Brand” and the confirmation email arrives in English. It reads as unfinished, at the moment someone is deciding whether to trust you with cash.
The field list is longer than people expect: name, descriptions, attribute names and their values, category names, filter labels, size charts, shipping copy, order templates, invoices and payment errors. Attribute values are missed most often: the platform shows the label as translated while the value passes through untouched.
Two details decide whether it looks right. Numerals and currency stay left to right inside a right-to-left paragraph, so a price has to be tested in place. And Arabic search needs normalisation for the forms of alef, for ta marbuta and for omitted diacritics, plus the Latin spellings buyers type for brand names, or your own search returns nothing for a product you stock. Product markup needs a variant per language: see schema.org/Product and Google’s product data documentation. We argue against machine translation every time, because native readers spot it and the place they spot it is the checkout.
WooCommerce, Shopify, Salla or Zid
There is no best platform, only a best fit for your catalogue, your market and whoever maintains it. We will talk you out of the wrong one.
| Decision | WooCommerce | Shopify | Salla or Zid |
|---|---|---|---|
| Catalogue size | Hundreds to many thousands of lines, given proper hosting | Thousands comfortably; complex variants hit platform limits | Strongest under a few hundred lines; deep catalogues get awkward |
| Arabic and right-to-left | Total control, and entirely your build’s responsibility | Theme-dependent; quality varies and usually needs custom work | Arabic-first by default, admin included, which matters if your team works in Arabic |
| Local payments | Any gateway you can get a merchant account for, integrated by us | Mada, Apple Pay and instalments through apps; cash on delivery built in | Mada, wallets, instalments and cash on delivery wired in from signup |
| Running cost | Hosting plus plugin licences. No revenue share, and no growth with sales | Monthly subscription in dollars plus app fees, rising with the plan | A monthly plan in riyals. Cheapest to run for a small Saudi catalogue |
| Ownership and exit | You hold the database and the code, and can move host at will | Products and customers export cleanly; theme, apps and custom logic do not | Product data exports; the storefront does not leave the platform |
| Who it suits | Content-heavy brands, odd pricing or shipping rules, teams that want the code | Teams who would rather pay for operations than run them, selling beyond the Gulf | Saudi-first sellers who want to trade quickly: Salla or Zid |
The honest summary: if your catalogue is small, Saudi-focused and speed matters more than control, a hosted Saudi platform beats anything we could build, and we will say so. If the store carries deep content, odd business rules or several languages, a self-hosted build pays for its flexibility. Between those poles it comes down to who maintains it on a Tuesday afternoon.
Addresses and phone numbers
The default address form is a Western artefact: street, city, state, postcode. In much of the Gulf two of those four are meaningless, and marking them required teaches buyers to type nonsense.
What each market uses is specific. The UAE has no delivery postcode: buyers give a building name, floor and flat with a landmark, and buildings carry a national geographic address number. Saudi Arabia has a full national addressing system, including a short code of four letters and four digits alongside building number, street, district and postcode, and couriers prefer the short code to the prose. Qatar runs on zone, street and building. Kuwait uses block, street, building, floor and flat.
So the form changes shape. Drop postcode and state, or make them optional. Split building and flat, because a courier reads them separately. Offer a map pin, the most valuable field you can collect. Take the phone with a country selector defaulting to the market, accept the local leading-zero form, store it in full international format, and give a second number its own field. Then close the loop: the address has to reach the courier’s label format, because many deliveries returned as “customer not available” are address failures wearing a different hat. Local detail sits on our Dubai, Jeddah, Doha and Kuwait City pages.
Returns, and what buyers expect
Under cash on delivery the return happens before the sale completes. The buyer opens the box at the door, decides, and hands it back if it is wrong. That is normal, and it is why photography, accurate sizing and honest Arabic descriptions are a logistics investment rather than a marketing one.
After delivery, exchange is often preferred to refund, and offering exchange or store credit first is normal here rather than obstructive, provided a refund stays genuinely available. The mechanics differ by method: a card refund goes back the way it came and takes days, while a cash order has no card to refund, so you need a bank transfer, an IBAN captured and a stated timeline.
Publish five things: the window, the condition the item must be in, who pays return shipping, how money comes back per payment method, and the same meaning in Arabic. The legal floor is not optional: Saudi e-commerce regulations give consumers a statutory right to withdraw within a set period, and UAE consumer protection law requires you to publish a policy and honour it, so “all sales final” survives contact with neither. Have the wording checked locally: we are a web studio, not your lawyer.
Migrating an existing store
Most of our store work is a move rather than a first build, and migrations go wrong in predictable places.
| What moves | What breaks if you skip it | How to verify after launch |
|---|---|---|
| Product and category URLs | Best-selling pages return 404 and their rankings go with them | Crawl the old sitemap; every URL must return one permanent redirect to a live, relevant page |
| Product data, variants, stock | Orphan variants, wrong prices, codes that no longer match the warehouse | Export both catalogues; compare code, price and stock line by line |
| Arabic content | Part of the catalogue silently falls back to English for a month | Filter for products with an empty Arabic name or description, and clear the list |
| Customer accounts | Everyone is locked out at once and blames the new store | Passwords rarely transfer. Send a reset invitation and watch first-login numbers |
| Order history | Support cannot answer “where is my order” and open returns stall | Spot-check twenty old orders, including one refund and one part-shipped order |
| Payment methods and stored tokens | Saved cards and recurring payments die at the cutover | Place one live order per method and refund it, then read the settlement report |
| Tax settings and invoice numbering | Non-compliant invoices, or a sequence that restarts | Download an invoice per market; check rate, registration details and sequence |
| Analytics and conversion tracking | A month of paid spend with no idea what it bought | Fire a test purchase; confirm it in analytics at the right value |
| Search engine access | A staging rule blocking indexing is carried over and the store disappears | Fetch robots.txt and the sitemap on the live domain, then check indexing in week one |
Two habits carry most of the risk. Keep the old store readable for a fortnight after cutover. And move in your quietest week, not the week before a promotion.
How a store build runs
01
Catalogue and payments audit
How many lines you sell, which payment methods you can actually get a merchant account for, who ships, and how returns come back.
02
Platform decision and fixed quote
A written recommendation with the trade-off stated plainly, a fixed price and a launch date. If a hosted Saudi platform beats a custom build, the quote says so.
03
Build, Arabic content, payment testing
You approve the home page and one product page before the rest is built. Arabic is written, not translated. Every method is tested with a real transaction.
04
Launch, then two weeks watching
We watch failed payments, abandoned checkouts and refused deliveries through the first fortnight. Domain, hosting and store stay in your name. Care plans run at AED 290 monthly or 2,190 yearly, and SAR 299 or 2,200, for updates, backups and monitoring. What we do not do is pack boxes or answer your WhatsApp: see pricing.
“Cash on delivery is credit you extend to a stranger. The confirmation message is the only underwriting you get.”
Related reading
If you are working out what to build and what it costs:
- Website costs in Qatar and Kuwait — the two markets with no VAT
- UI/UX design — the checkout and product page work described above
- A multilingual platform and a Dubai booking flow
- Riyadh and Abu Dhabi — city by city
- The blog — longer pieces on Gulf payments and Arabic content
Or tell us what you sell. A useful first call takes twenty minutes and needs only a rough product count.
E-commerce questions
Which platform will you build our store on?
WooCommerce, Shopify, Salla or Zid, chosen on catalogue size, operations, who maintains it and your budget. The recommendation comes in writing, and we have talked clients out of a custom build when a hosted platform suited better.
Can you migrate our existing store?
Yes, and it is most of what we do. Products, variants, customers, order history and rankings all move, with every old URL permanently redirected, using the checklist on this page.
Do you supply product photography and content?
Clients usually supply imagery, because it needs the stock in hand. We write descriptions, handle the Arabic and structure the catalogue and filters. For photography, hire locally.
How long does a store take to build?
Six to eight weeks for a focused catalogue, longer if it is large or customised, with a dated plan before work starts.
Do you handle Mada and cash on delivery?
Yes. Mada is essential for Saudi Arabia and we test it with a real transaction before launch. Cash on delivery gets a confirmation step before packing, a cap on order value and a rule for repeat refusers.
Will we lose rankings when we move platforms?
Not if the mapping is done properly: every old URL gets a permanent redirect to the closest live page. Recovery is usually weeks; after a careless migration it can take a season.
Who writes the Arabic product descriptions?
We can, or you can supply them. What we advise against is machine translation: native readers spot it immediately, and where they spot it is the checkout, the most expensive place to lose trust.
When should we sell on Salla or Zid instead of building our own store?
When your catalogue is small, your market is Saudi Arabia, and being live in weeks matters more than owning the code. Those platforms arrive with Mada, wallets, instalments and cash on delivery wired in, and an Arabic admin your team can use. Build your own when the store carries deep content, unusual pricing or shipping rules, several languages, or is the business rather than one channel of it.
Should we launch bilingual on day one, or add Arabic later?
Bilingual on day one if you can fund the content. Retrofitting Arabic costs more than building it in, because every field, filter, invoice and template has to be revisited and the URL structure usually changes, which means redirects. English-only is defensible while the catalogue is still moving. Launching machine-translated Arabic to look complete is the one option that does not work.
Do you have an office in the UAE or Saudi Arabia?
No. We are a studio in Istanbul working Gulf hours and we will not pretend otherwise. Everything runs on WhatsApp and scheduled calls, and the store stays in your name. If a local presence is a requirement, hire locally.
What does a Gulf online store cost with you?
Stores start at AED 16,500 in the UAE and SAR 16,800 in Saudi Arabia, custom builds from AED 29,000 or SAR 29,600, and simpler bilingual sites at AED 4,700. Prices are fixed in writing. VAT is 5% in the UAE, 15% in Saudi Arabia, none in Qatar or Kuwait.
Do we have to offer cash on delivery?
No, and some catalogues are better without it. If your average order is high, your margin thin or your courier coverage patchy, prepayment only is reasonable. Offering it unconfirmed, uncapped and with no handling fee usually costs more.
Can the store take orders through WhatsApp?
It can hand off from any product with the item pre-filled, and send confirmations and dispatch notices in the same thread. What we will not do is hide prices to force a conversation.
What returns policy should we publish?
One stating the window, the condition required, who pays return shipping, and how money comes back per payment method, including a bank transfer for cash orders. Publish the same meaning in Arabic, and have the wording checked by a local lawyer.
Ready to sell online properly?
Tell us what you sell and which markets. You will get a platform recommendation, a fixed price and a launch date.