Selling online in Iraq: the parts nobody mentions

The store is the easy half. Payment on delivery, addresses that are landmarks, Arabic search and the WhatsApp checkout are what decide whether it works.

Building the store is the part everyone quotes for. Product pages, a cart, a checkout, an admin panel — well understood, and by now close to a commodity. Then the store goes live in Iraq and meets the parts that were never in the quote: an order placed by voice note, an address that is a landmark and a phone number, a customer who pays only when the box is in his hands, and a driver reconciling cash on Thursday.

None of this is a reason not to sell online. It is a reason to design for the market you are actually in, rather than porting a template built for a market with postcodes and saved credit cards.

Cash on delivery is not a payment method. It is an operating model

Most Iraqi storefronts still take the majority of their orders as cash on delivery, alongside the wallets and bank apps that are steadily growing. Treating COD as "one more checkbox at checkout" is where the trouble starts, because COD changes the shape of the whole business:

  • An order is not revenue. It is an intention, and a meaningful percentage of intentions are refused at the door. Your dashboard has to separate placed, dispatched, delivered and collected — four different numbers that a template store usually collapses into one.
  • Stock is reserved for days. Goods leave the shelf on dispatch and only convert on delivery, so the stock model has to hold a middle state or you will oversell.
  • Cash comes back in a batch. Money arrives from the courier days later, in a lump, covering many orders — and it must reconcile against them individually, or you will never know which orders were actually paid.
  • Refusals cost real money. A refused delivery is a paid round trip. Track the rate per product, per city, and per customer; it is one of the most useful numbers in the business, and almost nobody measures it.

If your system cannot answer "how much cash is currently with drivers?", it is not an e-commerce system yet.

Addresses are directions, not fields

The address form on a stock store expects a street, a number and a postcode. Iraqi deliveries run on district, nearest landmark and a phone the driver will actually call. Force customers into the wrong fields and they will either abandon the checkout or fill it with punctuation.

What works: a province and district selector, a free-text field explicitly asking for the nearest landmark, a required phone number treated as the primary key of the delivery, and an optional map pin for those who will drop one. Then print all of it on the driver's manifest, because that is where it matters.

The checkout that actually converts is often WhatsApp

You can build a flawless three-step checkout and still watch a chunk of your customers screenshot a product and send it to your business number. That is not a design failure; it is a preference for talking to a person before money moves. Fight it and you lose the sale.

The better answer is to make both paths lead to the same system: a chat order gets created as a real order by whoever answers, with the same stock reservation, the same reference number, the same delivery record. What you must avoid is two parallel businesses — the website's orders in one place and the WhatsApp orders in a notebook. That gap is where stock accuracy goes to die.

Arabic-first is a product decision, not a translation task

A bilingual store built English-first and translated afterwards fails in specific, predictable ways:

  • Search. Arabic shoppers type without diacritics, with alef variants (أ, إ, ا) and with هـ/ة used interchangeably. Search that does not normalise these returns nothing for a product you definitely stock.
  • Product names. Half the catalogue will be typed in transliteration or in English by customers regardless. Both spellings need to find the item.
  • Numbers and units. Sizes, dimensions and prices need to read correctly inside right-to-left text — a detail that mangles constantly when it is an afterthought.
  • Layout. Mirroring is the beginning, not the end. We wrote separately about why true RTL is not a checkbox.

For Ghimar, the exclusive Somany agent in Iraq, the site was designed in Arabic first and mirrored into English — tiles are sold on look and feel, and a grid of thumbnails with translated labels would have flattened exactly what the showroom sells. Their team publishes products and articles themselves through a block-based page builder, which was the whole point: the developer should not be a dependency for content.

Photography decides the price you can charge

The single highest-return investment in an Iraqi online store is usually not the platform. It is a day with a photographer and consistent lighting. Customers cannot touch the product, cannot see it on a shelf, and are being asked to pay on delivery for something they have only seen on a phone. Sharp, consistent, honest images do more for conversion than any checkout optimisation.

Related, and just as unglamorous: write real product descriptions. Specifications, dimensions, what is in the box, what it is not. Every question your description answers is a WhatsApp message your team does not have to answer forty times.

The sale is the beginning of the relationship

A store that only sells is a store that has to buy every customer again. Almalakia sells electric bikes — a considered purchase where the web does the pre-sale storytelling — and runs the after-sale half in a companion mobile app: members, loyalty points issued and redeemed, orders, all visible on one dashboard instead of a notebook at the counter. That structure matters more the higher the ticket price.

What to settle before anyone writes code

  • Who packs the order, and where does the packing list print?
  • Who answers messages, on which number, during which hours?
  • What happens when a customer refuses delivery — who pays, and how is the stock returned?
  • How does cash from drivers get reconciled against individual orders?
  • Who writes and photographs new products, and how do they get published without a developer?
  • What is your return policy, in writing, in Arabic, on the site?

Answer those and the software becomes straightforward. Skip them and no platform will save you, because none of these are software problems — they are operations problems that software is asked to represent.

Build it once, run it yourself

We build storefronts and the CMS behind them as bilingual systems your own team operates after handover — catalogue, orders, content, and the integrations that connect them to payments, SMS and WhatsApp. And as with any project, it starts with a written scope, because "an online store" is exactly the kind of phrase that means five different projects to five different people.

More from the blog

All articles