Ecommerce website features: the counted list from 90 real store platforms
Most ecommerce website feature lists read like a platform's pricing page: product pages, cart, checkout, payments, shipping, discounts, reviews, wishlists, SEO, analytics, abandoned-cart emails, loyalty. Every item is real, and a list like that is almost useless, because it does not tell you which features a store like yours actually needs or what separates a working feature from one that only looks finished.
This list is counted. We went through 90 sources — platform documentation, open-source store software, store scripts and live admin demos — and tallied a capability only when a source directly showed it in use. For each feature you get the real version, the decorative one, and a test you can run on any store — yours, a platform trial, or something a developer or coding agent has built for you.
The first thing the count shows
Look at the numbers below and one thing stands out: nothing is universal. Even the most common capability appears in about half the sources. That is not because the other half are broken. It is because "online store" covers very different businesses — a showcase catalogue where customers enquire or pay on collection, a small shop with a handful of products, a store with thousands of variants and daily orders.
So the first decision is not which features to tick. It is what kind of store you are running. Customer accounts can wait for a showcase store; they cannot for a shop with repeat customers. A payment integration is a branch you choose, not a default. Get that decision right and the feature list shrinks to what you will use.
The nine, ranked
| Feature | Sources (of 90) | Share |
|---|---|---|
| Product variants and options | 49 | 54% |
| Customer accounts | 48 | 53% |
| Order management | 47 | 52% |
| Checkout and order creation | 44 | 49% |
| Payment gateway integration | 42 | 47% |
| Search, filter and sort | 42 | 47% |
| Shipping and fulfilment | 40 | 44% |
| Inventory tracking | 39 | 43% |
| Reviews and ratings | 35 | 39% |
1 · Product variants and options — 49 of 90
Size and colour, and every combination of them.
Decorative version: a dropdown that changes the label. Real version: each variant has its own SKU, price and stock, and an order records exactly which variant was bought at what price.
The test: buy a medium in blue, then change the price of the medium blue. Open the earlier order. It should still show the price the customer paid. If it shows the new price, orders are reading the live product instead of recording what was sold.
2 · Customer accounts — 48 of 90
Decorative version: a sign-up form. Real version: a profile with saved addresses and an order history the customer can see.
The test: save two addresses, place an order to the second, then look at your order history. For a showcase store this can wait; for a shop with repeat customers, it cannot.
3 · Order management — 47 of 90
Decorative version: an email to the owner for each order. Real version: an order list and an order page for staff, showing what was bought, for how much, where it is going and what state it is in.
The test: find an order using only the customer's email address, and check you can see the items, prices, address and status in one place.
4 · Checkout and order creation — 44 of 90
Decorative version: a checkout page that looks complete. Real version: completing checkout creates a durable order, or a record of handing the customer to the payment provider, that exists whatever happens next.
The test: complete a checkout, refresh the confirmation page and go back a step. There should be exactly one order in the admin, not zero and not two. A checkout that has only ever been tested with an empty cart has not been tested.
5 · Payment gateway integration — 42 of 90
A branch, chosen when you take card payments online.
Decorative version: a payment button that marks the order paid when the customer lands on a thank-you page. Real version: the order is marked paid when the payment provider confirms it, and a confirmation received twice is recorded once.
The test: pay for an order, then press Back and try to pay again. You should see the same paid order, not a second charge. Then start a payment and abandon it. The order should not be marked paid.
6 · Search, filter and sort — 42 of 90
Decorative version: a search box that matches product titles. Real version: filtering by the attributes customers actually choose on — size, colour, price — and sorting that behaves as labelled.
The test: filter to one size and sort by price. Check the results against your stock. Out-of-stock variants showing as available is an easy way for this to go wrong, and customers notice it at checkout.
7 · Shipping and fulfilment — 40 of 90
Decorative version: a single shipping charge added to everything. Real version: flat-rate delivery, local delivery and collection handled as separate options, with carrier integrations as a branch.
The test: choose collection at checkout. You should not be charged for delivery, and the order should say it is for collection.
8 · Inventory tracking — 39 of 90
Decorative version: a stock number that goes down. Real version: every change to stock is recorded as a movement, and stock is checked again at the moment the order is created, not only when the item went into the basket.
The test: put the last unit of a product in two baskets, in two browsers, and check out both. One should fail. If both succeed, you have sold something you do not have.
9 · Reviews and ratings — 35 of 90
Decorative version: stars on the product page. Real version: reviews tied to real orders and moderated before they appear.
The test: try to review a product you never bought. Whether that should be allowed is your decision — but it should be a decision, not an accident.
Which of the nine your store needs
The count says nothing is universal, so here is the same list read by kind of store. Reviews are optional for all four.
Single-product page. Needs checkout and order creation, order management, payment or an honest handoff, shipping and stock. Variants only if there is a size or two. Skip customer accounts and search.
Catalogue. Needs the other eight. Customer accounts earn their place once buyers come back, and search, filter and sort as soon as customers browse rather than land on one product.
Digital downloads. Needs checkout, orders and payment — access should follow a confirmed payment — and usually accounts, because buyers come back for their files. Skip shipping and stock. Variants rarely, and then only formats or licences.
Productised service. Needs checkout, orders, and payment or a deposit, because the order is the start of the work. Scheduling takes the place of shipping, and capacity the place of stock. Accounts help clients come back to book. Skip search.
Everything you can skip is a feature you can leave out without the store being incomplete, and leaving those out is most of the difference between a store that launches in weeks and one still being built a year later. Marketplaces and store-builder platforms are left off on purpose: they are different products, covered in ecommerce platform for small business.
What is deliberately not on the list
Discount codes, wishlists, abandoned-basket emails, gift cards, subscriptions, multi-currency and marketplace selling are all real features and all choices. Each brings its own rules: discounts have to recalculate tax on the reduced price; subscriptions bring renewals and failed payments; marketplaces bring sellers, payouts and commission. Add them when the store needs them, not because they appear on a comparison page.
The requirements list you can hand someone
- Store type decided first, in writing: [showcase / small shop / large catalogue].
- Each variant has its own SKU, price and stock. Orders record the variant and the price paid.
- [If repeat customers] Customer accounts with saved addresses and order history.
- An order list and order page for staff with items, prices, address and status.
- Checkout creates exactly one durable order, however the customer navigates.
- [If taking card payments] Orders are marked paid only on the payment provider's confirmation, and a repeated confirmation never creates a second charge.
- Filtering by real attributes, with out-of-stock variants shown as unavailable.
- Delivery, local delivery and collection as separate options.
- Stock changes recorded as movements, and stock checked again when the order is created.
- Branches chosen in writing: [reviews / discounts / subscriptions / gift cards / multi-currency].
Why AI-built stores miss these
Ask a coding agent for an online store and it will produce a convincing one quickly — product grid, product page, basket, checkout, order confirmation — because that is what online stores look like. What it will not add unprompted is what happens under the checkout button: a variant price recorded at the moment of sale, a payment recorded exactly once, stock checked again before the order is taken.
When three AI models built the same store specification, the differences between them were in this layer — charging exactly once, never selling the last item twice, and not counting cash on delivery as money received. The online store build-off write-up walks through each one, the scores are on the build-off board, and all three stores are live on the online store demo page.
If you are still choosing a platform rather than building, ecommerce platform for small business covers what matters at that size, free ecommerce platforms covers what the free plans hold back, and best ecommerce platform for SEO covers the search side.
If you are pointing a coding agent at this, the Online Store planner is the specification we hand ours: the store type locked first, and the checkout, stock and payment rules written in so an agent has to build them.