How to Sell Online in South Africa

Six things, in roughly this order. The sections below take them one at a time.

1. A platform to run the store. Shopify for most; Wix Studio if selling is a sideline.
2. A way to take payment locally. A card gateway, usually paired with instant EFT.
3. Products and content ready to load. Titles, images, variants, weights and box sizes.
4. Shipping and tax wired into checkout. A delivery company connected, and VAT worked out on the goods and the delivery together.
5. POPIA handled. Privacy policy, cookie consent, opt-in marketing, and a registered Information Officer.
6. A launch check before you go live. A real order placed from the customer's side, start to finish.

Take them in that order. Payments and delivery decide whether you have a business; design decides how it feels. Every store I have seen stall at ninety percent finished was waiting on one of these six, not on how it looked.

For a store whose main job is selling products, Shopify is usually the right call: checkout, inventory and payments are built for exactly that. Wix Studio suits business sites that sell lightly. Genuinely unusual requirements are an Audit conversation, not a platform guess.

The deciding question is not features, it is upkeep: who patches, backs up, and fixes it when it breaks. Owners choose on features and then live with maintenance, and every emergency I have been called into began as a maintenance job somebody deferred. The full comparison, including WooCommerce and a short diagnostic, is in Shopify vs Wix vs WooCommerce. Whichever you land on, it only starts mattering once it can take money.

Most South African stores offer two ways to pay, not one. Cards cover the customers who have one and will use it. Instant EFT covers the security-conscious: people who bank online daily but will not type a card number into a shop they have never bought from. Offer cards alone and you lose them at checkout, the most expensive place to lose anyone.

Sorted by what they let a customer do, the local options look like this.

• Cards. PayFast, Yoco, Peach Payments, Paystack.
• Instant EFT. Ozow specialises; PayFast and Paystack do both on one account. The customer pays from their banking app and enters no card number.
• Buy now, pay later. PayFlex pays you in full upfront while the customer spreads the cost.

None of these is an either/or. A checkout can carry several at once, and where there is no monthly fee, an unused option costs nothing.

Stripe is the one people ask about. It does not settle in rand here, and Paystack is the nearest equivalent: Stripe-owned, local, cards and EFT on one account (it is what I use for the studio).

You do not need a registered company to start. SnapScan and Zapper onboard sole proprietors on QR codes and payment links, enough to sell from a social account or an invoice. Yoco is strongest in person but plugs into Shopify, Wix and WooCommerce. All three want your ID and a South African bank account.

Which pairing suits you depends on how your customers actually pay, and you will not know that for a month or two. Start with one card gateway and one EFT option, then drop whichever goes unused. The payment gateways guide compares fees and settlement speed.

Two things shape your launch more than the integration does. Every provider verifies your identity before releasing money, and I have never had an integration take as long as that document check, so gather the paperwork early. Every provider also holds each payment for a short settlement period.

That settlement hold reads like a delay and works more like a buffer. A refund inside the window costs you a few clicks. After it, the same refund is money you have already counted going back out, with a fee on top and an afternoon spent arranging it. Fixing a mistake cheaply is usually what keeps the customer who made it.

If you are still working out how to trade legally, that comes before any of this: the register to sell online guide covers sole proprietor, Pty Ltd, and when VAT starts to matter. With a way to get paid settled, the store needs something to sell: product data and page copy in a state you can actually load.

Clean product data is what lets a catalogue import without errors and show up properly in search. Five fields do most of that work.

• Titles. What a customer would search for, not the name you use internally.
• SKUs. One unique code per version of a product, so stock counts stay attached to the right thing.
• Variants. Size, colour and material, named the same way across the whole range.
• Alt text. A line per image saying what it shows, for shoppers using a screen reader and for search.
• Weights and box sizes. These set what delivery costs, which the next section gets into.

This is the only item on the list you control outright, and the one I most often watch a build stall on. A finished store waiting on photographs is still a store you cannot open, and the platform subscription runs whether or not there is anything on it to sell.

The work travels further than the store, too. The same clean catalogue is what feeds Google Shopping and a Meta catalogue later, so a product loaded properly once is a product you do not redo for every channel you add. The product data formatting guide and the website content prep guide cover the fields and the page copy in turn.

Keeping it honest afterwards is a weekly habit rather than a project, and the managing prices and stock guide sets out the routine. Get the weights and box sizes right while you are in there, because the next step runs on them.

Getting a parcel to a customer involves two choices that are easy to mistake for one. The first is which company carries it. The second is whether you deal with those companies directly, or through one service that connects to several at once.

Your checkout can charge a flat delivery fee, or a live rate calculated for the actual parcel and address. I have yet to take over a store where the flat fee was still right. A live rate needs your delivery company to connect to your store platform: no integration, no live rate. Couriers publish their own connections, so this is not an aggregator-only feature, and the shipping and delivery guide covers them one by one.

That rate is calculated from two things you supply: the parcel's weight and its box size. In the stores I take over, accurate box size is the one usually missing. What that costs depends on the carrier: some bill on weight, some on dimensions, some on whichever is larger, and each measures a parcel its own way.

Size also decides whether a service is open to you at all: a locker parcel has to fit the locker, a per-sleeve service only works if your product fits the sleeve. Standard sizes pay off away from checkout too, making packing faster and stock and packaging predictable. Check how your carrier calculates before you standardise, so what checkout quotes is what it bills you. That settles delivery; one amount still goes on top.

That amount is VAT, and it applies to the delivery fee as well as the goods. If you are registered, the tax on an order is worked out on the two together, not the goods alone. Your store can only do that itself if it knows the delivery price at checkout, which is the live rate again. Fix that before an order ships at the wrong total. What remains is how the store looks after the personal details that customer handed over.

POPIA is South Africa's personal information law, and an online store sits squarely inside it. Checkout collects names, addresses and contact details; a newsletter signup collects more. Handling that lawfully is four pieces of work, not an open-ended obligation.

• A published privacy policy. What you collect, why, how long you keep it, and who else sees it.
• Working cookie consent. Nothing that tracks a visitor may run before they agree to it.
• Opt-in marketing. A box the customer ticks themselves, never one you tick for them.
• A registered Information Officer. By default that is you, registered with the Information Regulator.

The penalty everyone pictures is a fine. The one that actually arrives is a customer asking what you hold on them, and a store that cannot answer spends days reconstructing it from order emails. Every store I have taken over was missing at least one of the four, most often the consent banner, which was loading analytics before anyone had clicked anything.

Done once, these stay done, give or take a policy review when what you collect changes. None of it is legal advice from me, and a store handling unusual data should get some. The POPIA for online stores guide walks through each one. With the store lawful and the goods moving, the open question is what building it costs.

There is no honest fixed price to quote before scoping, because two stores selling the same number of products can involve very different amounts of work. Six things move the number.

• Catalogue size and variants. Fifty products in three sizes each is a hundred and fifty things to load, not fifty.
• Integrations. Every system the store has to talk to, whether that is accounting, a courier, or a marketplace feed, is its own piece of work.
• Migration. Moving an existing store means carrying over old addresses, orders and customer records, not just rebuilding the pages.
• Content readiness. Copy and photography you already have cost nothing to use.
• Custom features. Anything the platform will not do without being made to.
• Scope certainty. If nobody can yet say what the store has to do, that is an Audit rather than a quote, and half the Audit fee comes off the build if you go ahead within 30 days.

Readiness is the one driver you control outright. A catalogue that arrives as a clean spreadsheet, with weights, box sizes and image filenames matching the product codes, loads in one pass. The same catalogue arriving as photographs in a WhatsApp thread turns into a series of questions, and the questions are what take the time. The guide to what drives store scope goes through each driver in turn.

With the scope settled and the store built, one question is left, and it is the one people skip: whether the thing is actually ready to open.

Before going live, run payments, shipping, legal, content and the technical basics in one pass. Work the online store launch checklist, or use the readiness scorecard to see which gaps are still open.

Launch day is the start of the work, not the end of it, and getting your first sales picks up where this guide stops. The questions below are the ones I field most often before a store opens.