How to choose an ecommerce platform in 2026: a ranking by scenario, the criteria and when headless fits

What types of ecommerce platforms are there?

Ecommerce platforms fall into four types: SaaS, open source you host yourself, enterprise suites and headless engines you build your own store on. The types differ in who runs the store, how you pay and where the limits are.

The type matters more than the brand. Once you decide you need a rented platform, you compare Shopify with BigCommerce on prices and apps. Once you decide you need your own logic, you compare agencies and engines. Keep those two decisions apart.

Ecommerce platforms ranked by scenario

A ranking by scenario names the first choice for a typical sales situation and the point where that choice stops being enough. We do not award points, because a score without the context of your process compares features your store may never use.

The table does not replace a test on your own process, but it narrows the choice to one platform type. If your store fits two rows, pick the row with the harder process, because that process decides whether your team works in the store or around it.

Which criteria matter when choosing an ecommerce platform?

Six criteria decide the choice: your hardest process, the ERP integration, the 24-month cost, ownership of data and code, local payments and carriers, and who develops the store after launch. The feature list comes second, because every mature platform has a cart, variants and promotions.

1. Your hardest process. Pick the one order that causes the most trouble today (a discount outside the price list, a split shipment, an approval on the customer’s side) and walk it through the demo of every platform, from cart to invoice.
2. The ERP integration. Decide which system is the source of truth for prices, stock and customers, who maintains the connector and what happens to exceptions. We describe the scope of that work on our ecommerce integrations page, and the trade-offs in ERP and ecommerce integration.
3. The 24-month cost. The subscription after any discount, payment fees, apps, implementation, hosting and your team’s time. The starting price tells you little.
4. Ownership of data and code. Check that you can export products, customers and orders in full and whether the store’s code belongs to you. That decides the cost of your next migration.
5. Local payments and carriers. The payment methods and carriers your customers expect in each market. On global platforms, some of them only work through third-party apps.
6. Who develops the store after launch. A rented platform evolves on the vendor’s roadmap. Your own store evolves on yours, but it needs a team or a partner who stays after go-live.

What is headless ecommerce and when does it make sense?

Headless ecommerce is an architecture in which the storefront is a separate application, while products, prices, the cart and orders are handled by a backend that exposes an API. You can then develop, deploy and replace the storefront and the backend independently, and one backend can serve several storefronts.

Medusa is an example of a headless commerce engine: the Medusa deployment docs describe deploying the backend and the storefront separately, and the core is released under the MIT licence. We built Alufix on Medusa with a Next.js storefront this way. More on the engine itself in what Medusa.js is and when to build a store on it.

Headless makes sense in three situations:

• Several storefronts on one backend: a consumer store, a B2B portal for dealers, an app or a kiosk use the same products, prices and orders. We describe one such setup in a B2B customer portal for a dealer network.
• An unusual buying path: a configurator, a multi-step order with approval, or a cart a sales rep builds for the customer.
• Logic a theme cannot hold: prices calculated from the ERP in real time, customer-specific shipping rules, your own promotion rules.

Headless is the wrong choice when…

Headless is the wrong choice when you run one consumer store with a standard process, a small team without a developer and no plans for more channels. You then pay to build and run a second application, the storefront, without using the flexibility it gives you. In that case a SaaS platform with a ready-made theme is cheaper and faster to launch.

Choosing a platform for B2B: what to watch

In B2B, you choose an ecommerce platform by its price lists, trade terms and data exchange with the ERP, not by how the store looks. Ready-made platforms usually handle B2B with customer groups and paid add-ons. That works with a few buyer groups and starts to pinch when every customer has its own price list, a credit limit or an approval path.

How we build B2B stores with customer price lists, credit limits and ERP data is described on our B2B ecommerce development page. If you are weighing a custom build against an off-the-shelf product more broadly, the decision method is in custom vs off-the-shelf software.

How much does an ecommerce platform cost over 24 months?

The 24-month cost of an ecommerce platform is the subscription or licence, payment fees, apps and extensions, implementation, hosting and your team’s time. Entry prices differ by orders of magnitude, but as sales grow, the largest lines are usually fees that scale with revenue and people’s time, not the subscription.

A live store on Medusa built by Evelumo starts from €10,000 net with our live store in 30 days, and hosting with maintenance starts from €100 net a month. The full calculation with every line is in ecommerce website cost in 2026, and the Shopify side in detail in is Shopify worth it and Shopify Plus pricing. The calculator compares 24 months of Shopify with a custom store:

When a custom store is the wrong choice

A custom store on an open source engine is the wrong choice when a ready-made platform handles your hardest process without workarounds and the 24-month cost favours the ready-made option. In practice that covers a few situations:

• you sell to consumers with one price list and have no ERP that has to drive prices;
• you are still testing whether the product sells and launching within days matters most;
• nobody will develop the store after launch and there is no budget for a partner;
• the only reason to change is how the store looks, not how it works.

In these situations we recommend a ready-made platform, even when the conversation starts with a question about a custom store.

How to choose an ecommerce platform in 3 steps

You can choose an ecommerce platform in three steps: test your hardest process, price 24 months and ask the agency three questions about maintenance. The order matters, because the first step usually rules out half the options.

1. Test one process. Walk your hardest order through the demos of two or three platforms from one row of the ranking. Note what needs an app, a workaround or manual work.
2. Price 24 months. For the platforms that passed the test, add the subscription after any discount, payment fees, apps, implementation and your team’s time. Compare the same scope.
3. Ask three questions. Who maintains the ERP integration and what happens to exceptions? Do we own the code and the data, and how do we export them? Who develops the store after launch, and what does that cost per month?

If the answer is a custom store, see how a Medusa build works, and if you are leaving your current platform, start with ecommerce migration.