If we had to compress the answer into one sentence: open-source platforms suit more complex, deeply customized stores, while hosted (SaaS) platforms suit stores whose requirements are largely standard. That framing is still broadly true, but the landscape has shifted enough over the last decade that this decision now deserves a real framework rather than a one-liner — the wrong choice is expensive to live with and disruptive to reverse.
The two models, accurately described
Notable open-source platforms we work on include Adobe Commerce / Magento Open Source, WordPress with WooCommerce, OpenCart, and PrestaShop. The hosted platforms we work with most are Shopify and BigCommerce. A note for readers coming from older articles: "Magento Enterprise" and "Magento Community" were renamed Adobe Commerce and Magento Open Source after Adobe's 2018 acquisition, and platforms like Volusion and the discontinued Amazon Webstore are no longer ones we recommend for new builds. The category names changed; the underlying trade-off did not.
Hosted platforms are built so a merchant can launch and operate a store without being a developer. The vendor handles hosting, security patching, PCI compliance, uptime, and core upgrades — the parts of running a store that are invisible until they fail. The trade-off is a ceiling on customization: you build within the platform's APIs, theme system, and app ecosystem, and some deep changes are simply impossible because you cannot open the engine. You are trading flexibility for not having to think about infrastructure.
Open-source platforms give you the source code and, with it, effectively no customization ceiling and free licensing (Adobe Commerce being the paid, vendor-supported exception). The cost does not disappear; it moves. You become responsible for hosting, security, performance tuning, PCI compliance, and upgrades, and you need a developer who genuinely knows the platform available when something breaks — because there is no support line.
A decision framework you can actually apply
Run your real requirements through these questions instead of arguing the abstract:
- How custom is your product configuration? Standard products with simple variants? Hosted is more than enough. Dynamically configured products — the classic example being an engravable champagne glass that renders the buyer's typed text as a live preview — complex tiered B2B pricing, or bespoke checkout logic push you toward open source or a hosted platform's enterprise tier.
- What is your real operational capacity? Open source means someone owns patching, backups, and uptime forever. If no one on your team will reliably apply security updates within days of release, a hosted platform's managed security is not a limitation — it is a feature that prevents the breach you would otherwise eventually have.
- What is the honest total cost? Hosted platforms charge a predictable monthly fee plus possible transaction fees. Open source is "free" software with very real, recurring costs in hosting, development, and maintenance. Compare three-year totals including labor, not month-one license price — that is where most people get the math wrong.
- How fast do you need to launch? A hosted store can be live in days. A custom open-source build is a project measured in weeks or months. If time-to-market matters, that gap is decisive.
- Where is the business headed? Heavy ERP or PIM integration, a headless front end, multi-store or multi-region operations, and very large catalogs all push toward open source or an enterprise hosted tier. Build for where you will be in two years, not only where you are today.
The honest default
For most merchants, hosted is the better starting point: lower risk, lower operational burden, faster launch, and modern hosted platforms now cover the large majority of requirements that used to force an open-source build. The decision rule we actually use with clients is simple: choose open source only when a concrete, named requirement — not a vague desire for "flexibility" or "control" — genuinely cannot be met on a hosted platform. The worst and most common outcome we see is a store that picked open source for power it never used and now pays the maintenance and security tax on it every single month with nothing to show for it.
A worked example
Two stores, same revenue, opposite right answers. Store A sells roughly 400 standard SKUs of home goods, ships from one warehouse, runs no custom pricing, and has a two-person team with no developer. Store A should be on a hosted platform: every requirement it has is standard, it has no operational capacity to own security and uptime, and a managed platform removes an entire category of risk it cannot otherwise cover. Store B sells industrial equipment with configurable build-to-order products, customer-specific contract pricing, deep ERP integration for real-time stock and quoting, and has an in-house developer. Store B has concrete requirements a hosted platform cannot meet without painful workarounds, and the team to own the infrastructure — open source (or an enterprise tier) is the honest choice. The deciding factor was never company size or revenue; it was the specificity of the requirements and the capacity to operate the system. Run your own store through that lens before anything else.
This decision is not permanent
One thing that paralyzes people unnecessarily: the platform choice is reversible. Outgrowing a hosted plan, or finding an open-source build too costly to keep secure and updated, is common and survivable. We run ecommerce platform migrations in both directions regularly, and done properly — with a full URL map, 301 redirects, and preserved metadata — a migration moves the store without sacrificing the SEO equity it has built. Knowing you can change course later should make the initial decision less fraught, not more.
Frequently asked questions
Is open source always cheaper because the software is free? No, and assuming so is the most expensive mistake in this decision. Free licensing is more than offset by hosting, development, security, and ongoing maintenance costs that hosted platforms bundle into one predictable fee.
Can a hosted platform handle a large catalog and high traffic? Yes. Enterprise tiers of the major hosted platforms run very large, very high-volume stores. The old belief that scale automatically forces open source is outdated.
What about headless? Both models support a headless front end, but headless adds its own complexity and cost — treat it as a separate decision driven by a concrete need, not a default.
What if I genuinely cannot decide? Write down your must-have requirements as specific statements, then test whether a hosted platform can satisfy each one. Anything it cannot is your real argument for open source.
Still unsure which model fits your requirements? Tell us what you sell and how you sell it, and we will pressure-test your requirement list and recommend the platform that actually fits — not the one that happens to be trendiest this year.