En esta nota (10 secciones)
The question almost always comes after the first online sale: "do we leave this on the platform or build our own?".
Most of the time the right answer is to leave it where it is. It's worth understanding why, and also understanding the three specific cases where it isn't.
The three options, plainly
Tiendanube. The Argentine platform, today the most used in the country for selling online. The local side is solved: Argentine payment methods, shipping with local carriers, invoicing you can integrate.
Shopify. The global standard. More powerful, an enormous app ecosystem, better if you sell abroad. The local downside is that integrating with the Argentine world —payments, shipping, invoicing— depends on third-party apps and isn't always as direct.
A custom ecommerce. You build it. Anything is possible, and everything has to be built and maintained.
Why the platform almost always wins
An ecommerce isn't a catalog with a buy button. It's checkout, payment methods, shipping calculation, stock management, coupons, abandoned carts, transactional emails, an order dashboard, returns, and card data security.
The platforms already solved all of that, they keep it up to date, and they charge for it monthly. Rebuilding it costs far more than most people estimate, and the result, at best, is the same.
There's also an invisible cost: the platform improves on its own. Every so often a new feature shows up that you didn't ask for or pay for. Your own build doesn't improve unless you pay for it.
If your online business looks like any other store's —you show products, people buy them, you ship them— building custom is spending more to end up in the same place.
The three cases where building does pay off
1. When the price isn't a number
If the price depends on the customer, on volume, on a negotiated list, on a contract or on a formula, platforms get complicated fast. They're designed for retail selling at list price.
This is the most frequent case of all, and it's typical of B2B: each customer sees their own list, with their discount, their payment terms and the products they're allowed to buy.
2. When the sale is tied to your operation
If selling requires showing real stock across five warehouses, reserving production, calculating a delivery date based on the delivery route, or validating the customer's available credit, the store stopped being a store: it's a front end for your business system.
You can try it with integrations, and sometimes it works. But when the selling logic is your operating logic, keeping the two in sync costs more than having them together.
3. When the product isn't a product
Configurators, services with variables, quotes built step by step, made-to-measure products where the customer picks dimensions and materials. Platforms handle variants, not configurations.
What almost never justifies building
- "We want a unique design." Platforms allow custom themes. With front-end work on Tiendanube or Shopify you get to the same place for far less.
- "The per-sale commission is killing us." Do the math first: the build plus its yearly maintenance against what you pay in commission. The break-even point is usually further away than it looks.
- "We want to own the platform." Understandable, and it has a price. What is always worth guaranteeing is owning the data: customers, orders and history, exportable.
- "We need a feature it doesn't have." Check first whether an app solves it. It almost always exists and costs a fraction.
The middle path, the one used most and named least
It isn't platform or custom build. For most complex cases, what works is this:
The platform stays the store, and the custom work goes where the platform doesn't reach.
In practice: Tiendanube or Shopify facing the public, and your own build syncing stock and prices against your business system, assembling per-customer price lists, or handling the configurator and passing the result to the cart.
It's cheaper, faster, and it lets you keep what platforms do well —checkout, payments, security— without giving up your own logic.
How to decide in five minutes
Answer these three:
- Do all your customers see the same price? If not, the platform alone isn't enough.
- Can you sell without checking your business system in real time? If you can't, you need integration, no way around it.
- Is your product chosen or configured? If it's configured, the platform alone isn't enough.
Three comfortable "yes": stay on the platform and stop thinking about it. One or more "no": that doesn't mean building everything, it means you need your own layer on top of the platform.
The cost, for reference
A store on a platform with custom design and serious configuration starts around USD 1,500 to 5,000 to set up, plus the monthly fee and the commissions.
An integration between the platform and your business system —stock, prices, orders, invoicing— typically lands between USD 3,000 and 8,000, depending on how many systems and how many rules.
A full custom ecommerce, with all the selling logic your own, starts around USD 15,000 and goes up from there.
Those three numbers, put side by side, tend to organize the conversation better than any feature list.