How to start an online store. Platforms, payments, shipping and VAT — 2026 guide

Before you build: four decisions that come first

The most expensive way to start an online store is to start with the platform. The technology is the easy part. The hard part is answering four commercial questions that decide whether the store will ever pay for itself. Answer them in writing before you discuss a technical solution with anyone.

1. What do you sell, and why from you? A niche store with a clear reason to exist beats a broad “everything for everyone” catalogue almost every time, especially in a small market. In Estonia, with 1.3 million people, there is no room for a fifth identical store.
2. Who is the buyer and where do they come from? Search, Instagram, paid ads, or an existing customer list? A store with no traffic source is a website, not a sales channel.
3. Do the unit economics of a single order work? Cost of goods, shipping, packaging, payment fees, expected return rate and advertising cost per order. If your average basket is €25 and delivery costs €4, the margin disappears before the customer ever comes back.
4. Who runs the store after launch? Products, stock levels, orders, customer emails, updates. This is continuous work, not a one-off project.

With those answers in hand, the technical side becomes surprisingly simple — the rest of this guide walks through it step by step. If you would rather hand the whole build over, our e-commerce development page covers that route; this article is for people who want to understand the moving parts first.

Starting an online store comes down to five layers: business decisions → platform → payments → shipping → legal. None of them is optional, and the last three are local — that is where a real store diverges from any generic international checklist.

Choosing a platform: what the choice really is

For a small or mid-sized store the decision usually narrows to two options. Both handle European payment providers and carriers, and both scale far beyond a first launch.

• WooCommerce. An open-source store built on WordPress. Its strength is control: content, SEO, design and integrations are entirely yours, there is no licence fee, and connecting inventory or accounting software is nearly always possible. Hosting, speed and security updates become your responsibility — or your partner’s.
• Shopify. A hosted platform on a monthly subscription. Its strength is speed and peace of mind: hosting, security and updates are included, and a basic store can be live in an evening. In return you pay continuously, accept less design freedom, and face add-on costs for some apps and payment routes.

The short rule: if content and organic search are your main channel and you plan to publish articles and landing pages next to the shop, WooCommerce tends to win. If you want the fastest start with the least technical overhead and a simple catalogue, Shopify does. A detailed platform comparison deserves its own article — in practice, almost no store fails because it picked the “wrong” one of these two.

What genuinely matters when choosing: multilingual support (in the Baltics you usually need at least two languages), a properly fast mobile experience, and how easy the store is for you to run day to day. The same principles apply to ordinary website development — a store just raises the stakes, because here every second of load time costs orders directly.

Payments: global baseline, local habits

This is where generic guides mislead European sellers. Cards, Apple Pay and Google Pay are the 2026 baseline everywhere, but the method that converts is the one your market already trusts. In Estonia and across the Baltics, most shoppers do not reach for a card at all — they log into their bank and pay through a bank link.

• Local aggregators. In Estonia, Maksekeskus, Montonio and EveryPay bundle bank links from Swedbank, SEB, LHV and Coop Pank together with card payments and instalment partners under a single contract, with ready-made modules for WooCommerce and Shopify. They differ in how the money moves — some route it through their own account and pay out after a few days, others initiate the payment straight to yours — which matters for cash flow. Every market has its equivalent; find it before you design the checkout.
• Stripe. The sensible choice when a significant share of revenue comes from outside your home region, when you need subscriptions, or when you are building an international product. It will not give you local bank links, so on its own it is rarely enough for a domestic store.
• Fee structure. This varies by provider: some charge a fixed fee per bank-link transaction, others a percentage of turnover, while card payments are almost always percentage-based. With a low average basket the difference is visible by the end of the first month, so ask for the full fee table before you sign.

Whatever you choose, three things are no longer optional in 2026: wallet support (Apple Pay and Google Pay), strong customer authentication, and a refund flow you can actually operate without emailing support. Test all of it with a real transaction before launch, not after.

Shipping: match the habit, not the brochure

Delivery is a conversion feature, not a back-office detail. In the Baltics and Finland the default is the parcel locker: in Estonia, Omniva, DPD and SmartPosti (formerly Itella SmartPOST) cover the country and most buyers expect to see at least two of them at checkout. Courier delivery is an addition, not a replacement.

• Offer at least two networks. People have a habit and a favourite locker near home or work. A single carrier is the same kind of barrier as a missing payment method.
• Use a proper integration. Carrier and payment-provider plugins pull the locker list straight into checkout and generate labels without manual copying. Manual entry stops being viable at roughly thirty orders a month.
• Price delivery deliberately. Free shipping above a threshold raises average order value, but it has to be calculated rather than hoped for. Show the cost as early as possible — a surprise at checkout remains the number one reason baskets are abandoned.
• Plan returns before you need them. A 14-day withdrawal right means some goods will come back. Document the process and make it easy for the customer; that is cheaper than disputes and bad reviews.

Legal and tax: what the EU expects from a store

You do not need a law degree, but you do need to know which boxes must be ticked. This section is a general overview rather than legal advice — check the specifics for your business model with an accountant or lawyer.

VAT and OSS

Registration thresholds are national: in Estonia the obligation starts once taxable turnover passes €40,000 in a calendar year, and the standard rate has been 24% since 1 July 2025. For cross-border B2C sales inside the EU a common €10,000 annual threshold applies: above it you charge the customer country’s VAT rate, and the practical way to report that is OSS (one-stop-shop) — one registration and one quarterly return covering every member state. Build the rate into your pricing early; discovering it after you cross a threshold is an expensive way to learn.

Consumer rights

Distance selling to EU consumers carries a 14-day right of withdrawal with no reason required, and two years of seller liability for goods that do not match the contract. Terms of sale, the returns procedure, company details and contact information must be findable before the purchase, not after. Since 13 December 2024 the EU General Product Safety Regulation (GPSR) also requires clear manufacturer and safety information on product pages.

GDPR, cookies and accessibility

A store collects personal data by definition: name, address, order history. You need a privacy policy, a lawful cookie banner (marketing cookies only after consent) and a clear idea of where the data lives. Since 28 June 2025 the European Accessibility Act also applies to e-commerce, requiring most sellers to offer an accessible store — a checkout that works by keyboard, sufficient contrast, labelled form fields. Conformity is presumed through the harmonised standard EN 301 549, which points to WCAG 2.1 AA. Micro-enterprises (fewer than 10 staff and up to €2 million turnover) are exempt, but an accessible store is simply a better store: the same work improves conversion and SEO.

Budget: what the money actually goes on

There is no single answer, because a store is not a product but a set of parts. What follows is a general market benchmark for 2026; the ranges move with catalogue size, integrations and how much you do yourself.

• One-off build. A simple store on a ready-made theme with a small catalogue typically sits at the low four-figure end of the market; a project with custom design, several languages and real integrations runs into the tens of thousands. The biggest multiplier is not design but integrations: inventory, accounting, catalogue import.
• Running costs. Domain and hosting, platform subscription where applicable, monthly and per-transaction payment fees, carrier contracts and any paid modules. Budget for security updates and backups too — a cost that only becomes visible when it was skipped.
• Content. Photography, descriptions, translations. The most underestimated line in any store budget and the most common reason a launch slips by a month or two.
• Marketing. Ad spend and SEO. A store with no traffic source sells nothing, however good it looks.

A practical rule: plan the first year roughly half on building and half on running the store and driving traffic to it. The most common budgeting mistake is spending everything on the build and discovering on launch day that nobody is visiting.

Launch checklist

Work through this before you make the store public. Every line is something whose absence has cost real stores real orders.

1. Business. Company registered, bank account open, accountant lined up.
2. Products. Consistent photography, unique descriptions (not the supplier’s text), prices and stock levels correct.
3. Payments. The methods your market actually uses, plus cards, Apple Pay and Google Pay — all tested with a real transaction.
4. Shipping. At least two delivery options — in locker-first markets, two locker networks — with cost and delivery time visible in the basket.
5. Legal. Terms of sale, returns policy, privacy policy, cookie consent, company details in the footer.
6. Technical. SSL, mobile layout, speed (Core Web Vitals in the green), a real 404 page, backups.
7. Measurement. Analytics with e-commerce events (add to cart, checkout, purchase), Google Search Console, abandoned-cart email.
8. Content and trust. An About page, contacts with a real address, an FAQ, shipping and returns on their own pages.
9. Test. Place three complete orders using different payment methods and cancel one — that is how you find the bugs before a customer does.

One line deserves separate mention: search visibility. Structured data on product pages, unique descriptions, a sane category structure and speed decide whether the store is found at all — and in 2026 also whether AI answers (Google AI Overviews, ChatGPT, Perplexity) recommend your products. The logic is the same as everywhere else on the web: clear structure, quotable facts, a source worth trusting. We covered it in detail in our guide to generative engine optimization, and for a store it applies across the whole catalogue at once.

Seven common mistakes

• Choosing the platform before the business model. A month comparing WooCommerce and Shopify, zero hours calculating the economics of one order.
• Card payments only. The most expensive piece of technical short-sightedness in any market with strong local payment habits — you lose the buyer at the final step.
• Supplier product descriptions. The same text as ten competitors. Google sees no reason to prefer you, and an AI model has nothing distinctive to quote.
• Hidden shipping costs. The single biggest driver of abandoned baskets. Show the cost on the product page, or at the very least at the top of the basket.
• A slow mobile experience. Most store traffic arrives on a phone. Three seconds of loading means part of your audience never sees the product.
• No marketing budget. Launch day without a traffic source is a quiet day. Plan the channel before, not after.
• A store with no owner. Someone has to be responsible for stock, emails and updates. Without that, even a well-built store decays within six months.

If you worked through that list and concluded that several items are clearly somebody else’s job, that is a reasonable conclusion rather than a failure. A small merchant’s advantage is the product and the customer, not debugging a checkout.

Frequently asked questions