Start on a ready-made platform if demand for your product isn't proven yet. Consider a custom store when the platform's yearly cost passes what it would take to run your own, or when it can't do something basic to how you sell. A platform rents you a working store, hosting and updates included, while a custom store is software you own and maintain.
The first question at the start is whether people will buy. Spending a large budget on development before answering that is a gamble. After that, the decision is a calculation you can run. Below you'll find a decision table, a formula with a hypothetical example, and what to do before moving.
A decision table by order volume and integration
Find the row closest to your situation. The figures are our own rough starting points and no published standard, so adjust them for your margin and team size.
| Your situation | Closest choice | Why |
|---|---|---|
| Not selling yet, or under 100 orders a month | Ready-made platform | Your problem right now is proving demand |
| 100 to 1,000 orders, simple integration | Platform, with a fee review every six months | Usually enough if it isn't holding you back |
| 100 to 1,000 orders, staff re-entering data between store, accounts and stock | Platform plus an API integration | Connecting costs less than building a whole store |
| Over 1,000 orders with fees climbing steadily | Run the formula below | Moving may or may not pay off |
| A selling model platforms don't support: complex pricing, per-customer prices, contracts | Custom store, or a custom front end on a platform | Modifying becomes costlier than building |
| You own a custom store with no maintenance team | Review the decision before adding features | A store without updates becomes a security risk |
What does an ecommerce platform really cost?
The true cost is the subscription plus everything that piles up around it. Check these items on each platform's own page, since they vary and change:
- Monthly subscription: it usually rises with higher plans. See Zid's plans as the platform publishes them.
- A percentage of each transaction: the number that matters most as sales grow. For example, Shopify's billing page explains that it charges an extra fee when you use an outside payment provider, that the provider charges its own rates separately, and that the extra fee doesn't apply to manual methods like cash on delivery (checked November 2026). Confirm the rates and the payment methods available in your country.
- Add-on apps: each feature outside the basics may bring a separate subscription with its own billing cycle.
- Customization ceilings: some changes aren't possible, and you may only learn that when you need them.
- Team time: each order retyped into accounts or stock has a cost, even though it never appears on the platform's invoice.
The formula that compares both options
Compare the yearly cost of each option, and avoid setting a development quote against one month's subscription:
Platform per year = subscription + apps + platform percentage × sales + value of manual entry time.
Custom per year = build cost ÷ 3 years + hosting + maintenance and changes + security.
Leave payment gateway fees out, since you'd usually pay them either way. Many calculations miss this: a custom store removes the platform's percentage and leaves the payment commission in place.
A hypothetical example in generic currency units: a store selling 300,000 units a month, or 3,600,000 a year.
| Item (hypothetical) | Platform | Custom |
|---|---|---|
| Subscription and apps | 30,000 | 0 |
| Platform fee of 2% of sales | 72,000 | 0 |
| Build of 180,000 spread over 3 years | 0 | 60,000 |
| Hosting, maintenance, changes and security | 0 | 50,000 |
| Yearly total | 102,000 | 110,000 |
The two come out close here, so the sum alone doesn't justify moving, and the decision falls to the factors in the next section. If sales doubled, the platform fee would reach 144,000 and the balance would tip toward custom. If your percentage is zero, the platform wins by a wide margin. Use your own numbers.
Factors the sum doesn't show
Customer data ownership: your customer list and purchase history are among your most valuable assets. Ask up front: can you export all of it, and in what format? Shopify's help center, for instance, explains exporting customers to a CSV file. Test the same on any platform you choose.
Payment security: the PCI DSS requirements apply to every entity that stores, processes or transmits cardholder data. On a custom store that becomes your responsibility alongside the payment provider, so ask about a hosted payment page that keeps card data off your server.
Integration: if you run stock or accounting software, ask how the store will connect to it. A manual link is a problem that grows with each order.
Speed and measurement: find out how much control you get over page speed and tracking tools. Our article on improving website speed covers what to test, and setting up conversion tracking covers what you must be able to measure.
Invoices and tax: if you sell in Saudi Arabia or Egypt, check that your choice meets e-invoicing requirements. The Zakat, Tax and Customs Authority says taxpayers may choose any solution provider as long as it's compliant, and the Egyptian Tax Authority documents APIs for taxpayer systems. This is general information checked in November 2026, and it isn't tax advice.
Signs it's time to move
- Yearly fees exceed the custom cost by a clear and steady gap.
- You pay for several apps to get one job done.
- Your team re-enters data between the store and company systems.
- You need something central to how you sell and can't build it.
- The store is slow and you have no way to find out why or fix it.
One sign calls for a fix inside the platform first. Three together make moving worth serious study. The same logic applies to all company software, so also read how to decide between off-the-shelf and custom software.
If you move: three things you can't postpone
- Old URLs: Google's guidance recommends mapping each old URL to the right new one with a permanent redirect such as 301, keeping redirects for at least a year, and avoiding sending many URLs to the home page.
- Order and customer history: customers expect to find past orders, and your team may need them for returns.
- A parallel period: keep the old store open until the new one is taking real orders and payments, then test tax and shipping.
Next step
Pull the last 12 months of invoices, fill in the formula with your numbers, and find your row in the decision table. If you want to launch on a suitable platform, or build a custom store once the numbers justify it, take a look at PlanKit's e-commerce solutions service.