Which e-commerce platform should you build on?
The right platform depends on catalogue size, how much of checkout you need to control, and who maintains it on Monday. Onpolar builds on Shopify, WooCommerce, Adobe Commerce and BigCommerce, goes headless when the theme layer becomes the constraint, and builds custom when no platform fits. Migrations between them included.
- Platforms
- Shopify · WooCommerce · Adobe Commerce · BigCommerce
- Typical timeline
- 3–6 weeks for a storefront
- Migrations
- Full redirect map, rankings preserved
- Built this way
- Kullakelux · Eesti E-pood
The problem
The platform question usually gets answered backwards
Most stores end up on whichever platform the agency already knew. That holds until the catalogue grows, or checkout needs a rule the theme cannot express, or the app subscriptions quietly start costing more than a developer would have.
The other failure is the migration nobody planned. Moving platforms without a redirect map is how a store loses rankings it spent three years earning, in a fortnight.
We start from the catalogue, the checkout rules and who has to maintain it. Sometimes the answer is Shopify. Sometimes the answer is that you do not need a platform at all.
What you get
What we build for a store
Storefront build or rebuild
Theme-level on Shopify or WooCommerce where that is genuinely enough, or a custom frontend where the theme layer is what is holding you back.
Platform migration
Catalogue, customers, orders and content moved with a complete redirect map, so the rankings arrive along with the products.
Headless commerce
A storefront that renders independently of the commerce backend, for when speed, design or multi-market routing has outgrown the template language.
Checkout and payments
Local payment methods, tax and VAT rules per market, and the checkout customisation your platform actually permits — stated honestly up front.
Catalogue and inventory
Variant structures that survive growth, bulk import pipelines, and stock synced with whichever system holds the truth.
Multi-market and multi-currency
Localised slugs, currency and tax per region, hreflang naming one authoritative locale. Kullakelux runs three languages over a one-of-one catalogue.
Product SEO and GEO
Product and collection schema, crawlable faceted navigation, and product facts written where an assistant can quote them.
ERP and back-office integration
Orders, invoices and stock wired to the system that runs the business, so nobody re-types an order into a second tool.
How it works
How a store gets built, or moved
- 01Week 1
Platform decision, written down
Catalogue, checkout rules, integrations and maintenance reality scored against the options. You get the reasoning, not just the recommendation.
- 02Weeks 1–2
Data and redirect map
On a migration this comes before any design work: every existing URL mapped to its destination, and the catalogue export validated rather than assumed.
- 03Weeks 2–5
Build the storefront
Theme, custom frontend or headless, with product schema and localisation designed in rather than retrofitted after launch.
- 04Week 6
Launch and watch the rankings
Redirects verified in production, Search Console watched through the transition, and the first weeks of data actually read.
Proof of work
How this looks in production
One store built on Shopify because the platform was the right answer, and one agency platform taken off Wix because it was not.
Kullakelux — A Trilingual Shopify Storefront That Sells Trust Before It Sells Jewellery
Kullakelux restores pre-owned gold and diamonds from a counter in Tartu. We built the trilingual Shopify storefront that makes a second-hand ring feel safer to buy online than a new one.
Eesti E-pood — From Wix to a Trilingual SEO + GEO Engine for a Shopify Agency
We took Eesti E-pood off Wix in phase one, then rebuilt it as a trilingual 84-page platform — a WebGL commerce journey over an SEO + GEO engine built around one crown keyword and written to be cited by AI assistants.
Not sure which platform you should be on?
Bring the catalogue, the checkout rules and whatever is currently annoying you. Thirty minutes is usually enough to narrow it to one.
Platform fit
When each platform is the right answer
No platform is best. These are the conditions under which each one stops being the constraint — and the point at which it becomes one.
| Platform | Where it wins | Where it runs out |
|---|---|---|
| Shopify | Fastest route to a reliable store; hosted checkout and payments; deep app ecosystem | Checkout logic is limited outside Plus; app subscriptions compound; the theme layer caps frontend ambition |
| WooCommerce | Full control on top of WordPress; content and commerce in one place; no platform transaction fee | You own hosting, performance and security; plugin sprawl becomes the maintenance bill |
| Adobe Commerce | Complex B2B pricing, multi-store hierarchies, very large catalogues | Licence and hosting cost, and it needs people who know it to stay healthy |
| BigCommerce | Multi-channel selling with no transaction fee and generous native features | Smaller ecosystem than Shopify; the template language is a real ceiling |
| Headless | Storefront speed and design freedom; several markets or brands from one backend | Two systems to maintain, and content preview costs real engineering |
| Fully custom | Business rules no platform models — rentals, quoting, marketplaces, one-of-one stock | You are buying everything, including the parts a platform would have handed you free |
Kullakelux runs on Shopify because a one-of-one jewellery catalogue needed trust signals on every listing, not a custom cart. The platform stopped being the interesting problem.
FAQ
Questions we get asked
Which platform do you recommend?
Whichever survives your catalogue, your checkout rules and your maintenance reality. For most stores under a few thousand SKUs with standard checkout, Shopify wins on time-to-live. For content-led stores already on WordPress, WooCommerce avoids running two systems. We write the reasoning down so you can disagree with it.
Can you migrate us without losing our SEO?
That is the part we plan first. Every existing URL is mapped to a destination before design starts, redirects are verified in production, and Search Console is watched through the transition. Eesti E-pood came off Wix in two weeks on that basis.
What is headless commerce, and do we need it?
The storefront renders separately from the commerce backend and talks to it over an API. You need it when the theme layer is the thing holding you back — page speed, a design the template cannot express, or several markets from one catalogue. If none of those hurt yet, it is cost without return.
Do you build custom e-commerce from scratch?
When the business rules are not really a shop. Rentals, quoting, marketplaces, made-to-order, one-of-one inventory. Those we build on Go and PostgreSQL. If your rules are actually a shop, a platform is cheaper and we will tell you so.
How long does a storefront take?
Three to six weeks on an existing platform, depending on catalogue complexity and how many integrations are in scope. Migrations depend on data quality more than page count — a clean export moves fast, a decade of inconsistent product data does not.
Who maintains it afterwards?
You, us, or nobody — and it is worth deciding before the platform is chosen. A store your team can edit is a materially different build from one that needs a developer for a price change.
Can you connect the store to our ERP or accounting system?
Yes, and it is usually where the real time saving is. Orders, invoices, stock and customers flowing into the system that already runs the business. 3DCraft and Eesti E-pood both route theirs into ERPFlow.ai.
Does product SEO change now that people ask AI instead?
Product facts are exactly what assistants get asked for — materials, dimensions, compatibility, availability. Stores that state those plainly in served HTML with product schema get quoted. Stores that render them into a JavaScript tab do not.
Start an e-commerce project
Tell us what you sell, where you sell it and what is in the way. We reply within one business day.