Shopify vs a custom store: when each one wins

This comparison is usually written by someone who sells one of the two, which makes most of it useless. We build both. We have no preference and no partner commission either way, so the honest answer is available to us, and the honest answer starts with something a custom development studio is not supposed to say.

Shopify wins for most stores, most of the time, and the number of businesses that genuinely need a custom storefront is much smaller than the number being sold one. That is the headline. The rest of this is about the specific cases where it stops being true, because they are real and they are worth recognising early.

Where Shopify wins, and it often does

• You are selling within days, not months. Nothing custom competes with that, and time to first sale genuinely matters.
• Payments, tax and fraud are handled. Card handling, regional tax rules, chargeback tooling. This is a large amount of unglamorous work that you neither build nor maintain.
• It stays up during a spike. Getting a custom store through a launch or a seasonal peak is a real engineering problem that you simply do not have here.
• Someone else patches it. Security updates, platform maintenance, compliance changes to the checkout, all handled without you thinking about it.
• The ecosystem is genuinely deep. Whatever you need, something exists, and it usually works.
• Staff already know it. You can hire someone who has used it, or find a freelancer for an afternoon's change. Try that with a bespoke storefront.

That is a strong list, and the money you save by not building any of it is better spent on getting people to the store in the first place. Most stores that struggle do not have a platform problem, they have a demand problem, and no amount of custom development fixes that.

Where it starts to hurt

The friction is rarely one dramatic limitation. It arrives as a pattern: three or four things that each have an awkward workaround, and the workarounds start interacting.

• The app bill stops being trivial. Each one is modest, twelve of them are not, and several are charging per order so they grow exactly when you are doing well.
• Apps overlapping. Two doing similar jobs because neither quite fits, both loading on every page, both needing attention when either updates.
• Checkout is the wall. How much you can change depends on your plan tier, and if your selling model needs something unusual at that step, this is the constraint that does not bend.
• Your data lives somewhere you cannot fully query. Getting a real answer out of it means exporting into something else, and that pipeline becomes a thing you maintain.
• Variant limits and product shape. Made-to-measure, configured products, anything where a customer specifies rather than selects, tends to be fought rather than modelled.

One or two of these is normal and is not a reason to move. If you are living with several at once, outgrown Shopify works through whether what you have is a genuine platform limit, a missing app, or a setup problem, which is the distinction that saves the most money.

The four triggers that actually justify custom

In our experience these are the only ones that reliably survive contact with a spreadsheet. Everything else has a cheaper answer.

1. Trade and wholesale pricing that is genuinely per customer. Not a discount tier, but negotiated prices per account, agreed quantity breaks, and customers who must not see each other's terms. Platform features have improved here, and they still assume more uniformity than most wholesalers have. This is the trigger we see most, and B2B wholesale ordering covers it in detail.
2. Stock that is not a simple number. Serialised items, batches with expiry, stock held across branches with rules about which can fulfil what, made-to-order with capacity limits. If availability is a calculation rather than a count, you will fight the platform forever.
3. A checkout that is not a checkout. Deposits and staged payments, quote-then-approve, purchase orders and credit accounts, delivery slots with real constraints, price that depends on configuration. The further you get from card-now, the more this bites.
4. Integration is the actual product. When the store is really a front end onto an ERP or a production system, and the interesting logic lives there. At that point you are not buying a store, you are buying a way to expose one, and building the front end is often simpler than making a platform pretend.

Notice what is not on that list: design, speed, and wanting to own it. Design is solvable with a custom theme. Speed is usually a theme and app problem rather than a platform one. And ownership is a real value, but on its own it rarely justifies the cost difference.

The hybrid most people should consider first

This is the option that gets skipped, and it resolves a large share of the cases above. You keep the platform for what it is genuinely good at, and build only the piece that does not fit.

• A custom storefront on the platform's commerce, so you own the entire customer experience while payments, tax and orders stay where they are.
• A private app for your specific logic, replacing three ill-fitting public apps with one thing that does exactly what you need.
• A separate trade portal alongside the retail store, sharing products and stock. Often far cheaper than making one system serve two very different buyers.
• A proper data pipeline out, so reporting happens somewhere you can actually ask questions, without moving the store at all. That is our data pipelines work.

The hybrid is usually a third to a half the cost of a full migration, keeps the operational safety net, and can be reversed. If someone has quoted you a full custom build without first asking whether the hybrid would do, ask them why.

Shopify vs custom ecommerce, compared honestly

What migration actually involves

If you do move, the build is not the risky part. The risky part is that your existing store has search visibility, and a migration can lose it.

• Moves cleanly: products, variants, images, customers, historical orders, content. Careful work rather than difficult work.
• Needs real attention: every old address mapped to its new one. Skipping this is how businesses lose the traffic that was paying for the store. Expect some ranking movement even when it is done well, and be wary of anyone who promises none.
• Does not move: anything living inside an app. Reviews, loyalty balances, and above all recurring subscriptions, where customers often have to re-authorise payment. If you sell subscriptions, resolve this before committing to anything else.
• Gets rebuilt: the theme and every piece of custom work in it.

Budget for the fortnight after launch as part of the project rather than as a surprise. Something always needs attention, and the teams that plan for it have a calm week instead of a bad one.

How to decide, in one afternoon

The decision usually stalls because it is being argued in the abstract. It becomes tractable when you make it concrete, and this is the exercise we run with clients before quoting anything.

1. Write down the specific blocker in one sentence, without the word flexible and without the word scale. If you cannot, the platform is not your problem yet.
2. Add up twelve months of app subscriptions, including anything charged per order at your current volume rather than last year's. People are routinely surprised by this number, and it is the honest baseline for any comparison.
3. Count the manual work. Hours per week spent correcting orders, re-entering data, or working around something the platform will not do. Multiply by fifty. That is the invisible cost of staying, and it is usually larger than anyone assumes.
4. Ask whether an app already solves it. Search properly, not the two you have heard of. A surprising share of blockers have a twenty-a-month answer that nobody looked for.
5. Price the hybrid before the migration. If nobody has quoted you the middle path, you are comparing two options when there are three.
6. Check the subscription question early. If you sell on a recurring basis, find out what happens to those agreements before you evaluate anything else, because it can decide the whole matter on its own.

If steps two and three come to less than the cost of a hybrid spread over three years, stay and stop thinking about it. If they comfortably exceed it, you have a business case rather than a frustration, and the conversation can move on to which route.

Not for you if

A useful way to decide: write down the specific thing the platform will not let you do, in one sentence, without using the word flexible. If you cannot, you do not have a platform problem yet. If you can, and it is one of the four triggers, you have your answer and the conversation becomes about hybrid or full rather than whether to move at all.