What Shopify Is Bad At: Six Honest Limitations

Everything that follows is written by people who think Shopify is the best commerce platform a small merchant can rent. That is exactly why the limitations are worth setting out plainly: a tool this good gets bought for jobs it was never designed to do, and the disappointment always arrives after the catalogue has been uploaded.

Six things Shopify genuinely handles badly, and who should take each one seriously.

1. URLs and site architecture

Shopify hardcodes /products/, /collections/, /pages/ and /blogs/ into its router, and you cannot remove them. There is no setting, no theme edit and no app that changes it.

That means no hierarchical category paths — you cannot have /clothing/shirts/blue-tee, and a product cannot sit at the root of your domain. It also means Shopify generates a distinct URL for a product in each collection it belongs to, producing duplicate paths for identical content. Canonical tags contain most of the damage, but internal links can still emit the collection-scoped versions unless the theme is edited.

Anyone arriving from a platform with free URL control finds this genuinely irritating, and anyone whose strategy depends on a specific information architecture should understand it as a hard constraint. The detail and the workarounds are in the Shopify URL structure problem.

2. Content and editorial

Shopify's blogging tools are adequate and no more. The editor is basic, the taxonomy is thin, and the content model assumes that anything you publish is adjacent to a product rather than a reason for the site to exist.

For a store that publishes a care guide occasionally, this is fine. For a brand whose growth comes from editorial — long-form guides, a genuine publication, content as the primary acquisition channel — you will spend the whole time pushing against a system built around a catalogue. Plenty of such brands end up running content on a separate platform and treating Shopify as the checkout, which works but is two systems to maintain.

3. Checkout, beyond what the extensions allow

checkout.liquid is gone. Deprecated for the core steps in August 2024, retired for Plus thank-you and order-status pages in August 2025, with the non-Plus cutoff on 26 August 2026 now behind us. Shopify Scripts followed a separate track that ended on 30 June 2026.

The replacement, checkout extensibility, is more maintainable and considerably less free. You are working within defined extension points rather than editing a template, and several of the most powerful Functions — payment customisation, delivery customisation, order routing, cart-transform — are reserved for Plus.

For most merchants this is a non-issue and arguably an upgrade, since a hosted checkout that Shopify maintains is safer than one you have modified. For a merchant with genuinely unusual checkout requirements and no appetite for Plus pricing, it is a wall.

4. Cost predictability

Not cost — cost *predictability*. Shopify's plan prices are promotional and vary by country and session; the processing rate varies by card type and issuer country; the app bill varies by whatever accumulated during past launches. A merchant cannot easily answer what will this cost me, which is a strange weakness in a product otherwise so well made.

The fix is doing the arithmetic yourself rather than reading a pricing page, which is what Shopify pricing, plan by plan exists for. But it should not be the merchant's job, and a platform this dominant could afford to be plainer.

5. Owning your design

Themes are Liquid, and Liquid runs on Shopify. Content and product data export cleanly, which is the important part, but a storefront built over three years of theme customisation does not travel. Leaving means rebuilding the front end.

This is true of every hosted platform and is not a scandal. It is worth pricing honestly at the start, because merchants tend to discover it at exactly the moment they are least able to absorb the work.

6. Very small stores

Below roughly $1,143 a month in sales, our own crossover arithmetic says the fixed cost of a real Shopify plan is working against you. Add a theme and two apps and a hobby seller can easily be paying more in platform costs than the margin on what they sell.

Shopify is not built for that merchant and does not pretend to be — Starter exists for exactly this case. But a great deal of marketing encourages very small sellers onto full plans, and the honest advice is to stay on Starter, or on a marketplace, until the volume argues otherwise. The reasoning is in our Shopify Starter plan review.

The related trap is customisation debt. Every theme modification made over three years is work that has to be redone elsewhere, so the longer a store runs, the more expensive leaving becomes — which is precisely the dynamic that keeps merchants on platforms they have outgrown. Keeping customisation deliberate rather than incidental is the only real defence, and almost nobody does it.

What it is still excellent at

Selling physical products reliably at scale, without a developer. A checkout that works, that you never maintain, and that handles wallets and fraud without asking you anything. One stock pool across a website and a counter. And an ecosystem where almost any requirement has an existing answer.

None of the six limitations above touches that core. They matter when you are asking Shopify to be a publishing platform, a bespoke checkout, or a cheap option for a very small business — three things it is not. The balanced version of all this sits in our full Shopify review.

The useful way to read a list like this is as a specification rather than a warning. Every platform is bad at something, and the ones that claim otherwise are simply less honest about where the edges sit. What matters is whether the edges fall anywhere near your business. For a merchant selling physical goods, organising the catalogue in a way collections can express, and treating the checkout as plumbing rather than canvas, every constraint above is theoretical. For a content-led brand with a nested range and bespoke checkout requirements, four of the six bite immediately.

Work out which of those you are before the catalogue is uploaded, not after. That is the only decision here that is genuinely hard to reverse.