Starting an Online Store: Operations Come Before Themes

An online store fails more often from operations than from an ugly theme. Shipping rules you cannot honor, inventory you cannot count, returns you did not write down, and a payment method you cannot reconcile will sink a beautiful catalog. Pick the theme after you can describe a real order from click to doorstep (or to inbox, if you sell files).

If you already know the operations and you are choosing Shopify versus custom PHP, jump to the e-commerce website development guide. If you do not, stay here.

Write the offer as a fulfillment contract

Before a designer opens Figma, answer:

• What exactly ships or unlocks when the customer pays?
• Who picks, packs, or generates it, and in how many days you will actually commit to?
• Which countries you will serve on day one, and which you will refuse?
• What a return or refund looks like, including who pays postage?
• What happens when an item is out of stock after the order is placed?

Put the honest version of those answers on the site later. If you cannot answer them now, you are not choosing a theme; you are still designing a business. Ads will only scale the confusion.

Inventory and suppliers

Count stock in one system of record. If a supplier ships for you, decide whether their spreadsheet, their portal, or your store is allowed to say “in stock.” Two sources of truth sell the same unit twice. For made-to-order goods, sell lead time as a product fact, not a surprise in a shipping email.

SKUs need names a warehouse human can pick at 6 p.m. “Ocean / small / left” in the admin is better than a poetic title that does not match the bin label. Photographs should match the SKU you will send. That is a returns problem more than a branding problem.

Reorders: know minimum order quantities and how long the ocean or the printer takes. A store that looks live but restocks in 40 days needs a preorder mode, not a fake quantity of 999.

Shipping, tax, and the total the customer sees

Carrier rates, packing weight, and remote-area surcharges should be modeled before you promise free shipping. Free shipping is a discount you fund; it is not a carrier product. If you only ship inside one country at launch, say so in the cart, not in a FAQ nobody opens.

Tax: decide whether prices are tax-inclusive, which jurisdictions you must collect in, and whether a marketplace will collect for you. Do not launch a global checkout that silently under-collects. You can expand countries after the first corridor works.

Packaging is part of the unit economics. A fragile product in a pretty box that does not survive the carrier is a review, not a lifestyle photo.

Payments before pixels

Know how you will take money: cards, wallets, invoices, cash on delivery (if you truly can handle COD). Underwriting needs a site that describes the goods, a legal entity if the processor wants one, and a payout account. The practical path is in how to accept payments online.

Test a real refund in test mode and a small live refund after go-live. Stores that cannot refund become social-media stories. Connect payout reports to a bookkeeping habit from week one, even if week one is three orders.

Customer communication

Transactional email is operations: order received, payment failed, shipped (with a real tracking number), delayed, refunded. Use a domain you control (SPF/DKIM) so Gmail does not bury the tracking number. Support should have a path that is not only Instagram DMs. Instagram is not an order database.

If you sell across languages, decide which language the packing slip and the carrier label use. Support in a language you cannot staff is a promise you will break.

Legal pages that match the warehouse

Terms, refunds, and privacy should describe what you do. Copy-pasting a US-only returns policy when you cannot accept returns from abroad is how you lose disputes. If you are a non-resident with a US LLC, the site’s legal name should match the entity you will put on Stripe. Formation and processors are still separate approvals.

Then, and only then, the storefront

When the above is true, choose architecture (hosted, custom, hybrid) and a theme that does not hide price, shipping, or size. Product pages need: price, variant availability, shipping expectation, and photographs that load quickly on a phone. Collection pages need filters that match how people shop your catalog (size, category), not a moodboard of unused tags.

A homepage video that autoplays is not a substitute for a working search box if you have more than a handful of SKUs. Launch with a catalog you can photograph well. Expanding SKU count is cheaper than reshooting a lie.

A launch sequence you can actually follow

1. Fulfillment and returns on paper.
2. SKU list and stock source of truth.
3. Shipping countries and tax posture.
4. Processor application and checkout test.
5. Transactional email and a support inbox.
6. Product pages and a theme that does not fight mobile.
7. A small paid or organic test, then fix the first ten support questions in the copy.

Scriplit can build the site and integrate payments via web development and payment gateway integration once those operational answers exist. If you contact us with only a Pinterest board, we will send you back to this list. The stores that last are the ones that treated packing tape as part of the product.