Online Stores and Shopping Carts

Three costs, not one

Owners are quoted a single number for an online store. There are always three, they behave completely differently, and confusing them is the most expensive mistake in small-business ecommerce.

1. The platform fee. A fixed monthly subscription. Predictable, comparatively small, and the only number anybody puts side by side when choosing.
2. Payment processing. A percentage of every sale plus a flat fee per transaction. It scales with revenue, so it starts invisible and ends up, once the store works, the largest ongoing cost by a wide margin.
3. The work. Setup, product data entry, photography, tax configuration, shipping rules — and then, forever, order handling. This is the largest cost in year one and the part nobody budgets for, because it does not arrive as an invoice. It arrives as evenings.

The buying mistake follows from the structure: people optimize the first, ignore the second, and are quietly destroyed by the third. A store chosen because its monthly fee is slightly lower, on a plan that takes an extra slice of every sale, run by an owner who did not realize somebody must answer order emails on a Sunday, is the standard failure and it looks the same every time.

What a shopping cart actually is

Owners think of an online store as one thing. It is four separate jobs, and the split explains why quotes vary so much and why stores fail where they fail.

1. The catalog. Products, variants, prices, stock levels, photographs, descriptions. This is where the data entry lives.
2. The cart and checkout. Session state, shipping calculation, tax calculation, discounts. This is what people mean by "shopping cart", and on a modern platform it is largely solved.
3. Payment. Almost always handed to a processor through a hosted field or a redirect, so the site itself never touches the card number — which is why the compliance burden on a small hosted store is far smaller than owners fear.
4. Fulfillment. Picking, packing, labels, tracking emails, returns, refunds, and answering "where is my order". This is not software. It is a person, on a schedule, indefinitely.

Job four is where small stores fail. What breaks a store is the third week, when six orders sit unshipped because the person who was going to handle them had a busy weekend. One bad shipping experience produces a chargeback and a one-star review, and both cost more than the order was worth. So before building anything, answer out loud: who is the named person who packs orders, and what happens when they are on vacation? If there is no answer, the honest recommendation is not a smaller store. It is no store.

The fees that surprise people

Current rates change and we do not print them, because a rate typed into a paragraph is wrong within months. What does not change is the shape of the surprises.

The third-party gateway surcharge. Some hosted platforms charge an additional percentage of every sale if you do not use their own payment product, added on top of whatever your processor already charges. On an entry-level plan the combined effect can approach half again what you thought you were paying per order. It is disclosed — just not where people compare monthly fees.

Entry-plan transaction fees on website builders. Some let you sell on the cheapest plan but take a percentage of store sales that the higher plans do not. Once you sell regularly, the cheap plan is the expensive one.

Dispute fees. Every major processor charges a flat fee when a customer disputes a charge, usually whether or not you win. A business that ships slowly will meet this fee, and a written returns policy published before launch is the cheapest insurance against it.

Keep perspective, though. At the volumes a business here is likely to do in year one, the difference in card rates between the major processors amounts to a few hours of labor over a year. The differences that are large are the gateway surcharge and the entry-plan sales percentage.

Texas sales tax if you sell from Texas

This is not tax advice, rules and rates change, and a CPA is considerably cheaper than an audit. What follows is the structure of the obligation and where to read the real thing. The Texas Comptroller of Public Accounts is the primary source and is written more clearly than you would expect.

The rule people get backwards. If your business is located in Texas, you collect Texas sales tax on your taxable Texas sales. That is the starting point, and there is no small-business exemption from it.

The revenue threshold quoted in every forum thread is the remote seller safe harbor, and it applies to sellers located outside Texas. It is measured on gross revenue from taxable and nontaxable sales over the preceding twelve calendar months, and a remote seller who crosses it must begin collecting no later than the first day of the fourth month after the month the threshold was exceeded. There is also a single local use tax rate, published in the Texas Register each January, that remote sellers may elect instead of tracking every local jurisdiction — available to remote sellers only, not to a business operating from Kaufman or Henderson County.

Marketplaces versus your own website. A marketplace provider engaged in business in Texas must collect, report and remit state and local sales and use tax on sales made through its marketplace, and must certify to its sellers that it will. Absent that certification in writing, the seller should collect and remit until it has one. Marketplace sellers generally still need a Texas permit and still have to file returns on time, and records must be kept four years.

The consequence, plainly: selling on a large marketplace does not mean your tax is handled in general. Your own website is not a marketplace. On it you are the seller of record and you collect. That is the exact gap between "I already sell on Etsy" and "I launched my own store".

When a small local business should not build a store

This is the most useful section on the page and the one that costs us work. We would rather lose a store build than take money for something a payment link would have done.

A local service business. Plumber, electrician, septic, HVAC, dock builder, lawn care, tree service. There is nothing to put in a cart. What these businesses need is a way to take a deposit and get paid without waiting for a check: a payment link from a processor, or an emailed invoice. No monthly platform fee, processing only, an afternoon to set up.

A marina that sells four things. If the product range is a handful of items and the business is mostly service and slip rental, the right answer is a price list page and a phone number, or a price list page and a payment link. Four products do not need a catalog, a cart, an abandoned-cart email sequence and a shipping rules engine. They need to be findable and priced, and then somebody picks up the phone. That is not the lesser solution — it converts better, because there is no cart to abandon and no checkout to debug.

A boutique with three hundred items is a genuinely different conversation. Variants, sizes, stock that changes daily, seasonal turnover, photography as an ongoing production line rather than a one-off. At that scale a real platform earns its fee within weeks, because the alternative is manual and the manual version breaks. The dividing line is not "do you sell things". It is whether the catalog is large enough and changes often enough that maintaining it by hand would cost more than the platform, and whether stock genuinely has to be tracked. Under about twenty items sold mostly in person, a price list wins. Well above that, a store does.

Anyone already running a point-of-sale system in the shop. The major systems include an online store that already shares inventory with the register. Selling that business a second platform manufactures a synchronization problem that did not exist, and somebody ends up reconciling stock by hand forever.

A business that cannot pack and ship reliably yet. The store will generate chargebacks and one-star reviews faster than revenue. Fix the operation first.

A store nobody has agreed to own. "We should probably have one" is not a plan. If no named person has agreed to answer an order email on a Sunday evening, do not build it.

The mistakes that cost the most

In rough order of how much money they waste:

• Building a four-hundred-item catalog before proving anybody will buy online. Weeks of data entry against an untested hypothesis. Launch with the twenty items you expect to sell, then expand.
• Choosing on monthly fee alone and then meeting the gateway surcharge or the entry-plan sales percentage.
• Promising free shipping without modeling it. A low-priced item shipped at a loss is a business that loses money faster the better it does. Model the heaviest, farthest, cheapest order before publishing a policy.
• No written returns and refunds policy before launch. The first dispute gets improvised badly, under time pressure, with a fee attached to having the argument.
• Pasting manufacturer copy into every product description, producing pages identical to a thousand other stores.
• Deferring photography. A store with poor photographs does not convert on any platform. Highest return in the project, first thing cut.
• Ignoring shipping restrictions — alcohol, firearms and accessories, certain agricultural products, live plants across state lines, anything hazardous. Check before building, not after the first order.

Hosted platform or self-hosted

The choice comes down to where you would rather spend attention.

A hosted platform charges a subscription and handles servers, patching, checkout, compliance and upgrades. You get a system that works and cannot be heavily customized without effort. For most small local sellers this is right, and the reason is not technical: nobody in the business wants to think about the store's infrastructure, and on a hosted platform nobody has to.

A self-hosted store — the open-source route, usually a store plugin on a content management system — has no platform fee and no revenue share, which is where the word "free" comes from. The cost relocates into hosting, paid extensions and somebody keeping the stack patched. Store software is a far larger attack surface than a brochure site and it handles payments, so it cannot be left alone. Self-hosting is genuinely cheaper at volume and genuinely more expensive in attention.