Any e-commerce proposal that quotes a round number without asking about your operation is guessing. Not because the vendor is dishonest, but because the cost of an online store is not in the store — it is in everything the store has to talk to.
This post breaks the investment into the parts that actually exist and gives concrete ranges, so you can evaluate any proposal you receive, including ours.
The ranges, first
Numbers before explanation. These are the ranges we work with in the Brazilian market, for the profile we serve: a company that already sells, has no in-house development team, and runs a catalog of hundreds to a few thousand SKUs.
| Line | Range | When |
|---|---|---|
| Implementation (setup) | R$ 25,000 – 45,000 | R$ 25k is the entry point: B2C storefront, straightforward catalog, analytics configured. It climbs as capabilities like loyalty, B2B or subscription come in |
| Design and branding (optional) | R$ 5,000 – 50,000, averaging around R$ 15,000 | R$ 0 if you accept the standard visual kit. The average covers full identity plus special pages. The high end is a bespoke Figma with many templates |
| Integrations | R$ 0 to tens of thousands | Zero if the store talks to nothing. Each system is a mini-project. A legacy ERP with no documented API is the line that breaks budgets |
| Monthly retainer | R$ 3,000 – 5,500/month | Support, continuity, monitoring and evolution hours. Twelve-month contract |
| Typical full project | ~R$ 40,000 | Entry-level setup plus average design, without a heavy ERP integration |
One note on that last line: R$ 40k is an example, not a package. It works as an anchor to tell whether you are talking to the right vendor. A R$ 8k proposal for the same scope is leaving out one of these lines — almost always integration or ongoing work.
Now, what each line means.
The four cost lines
1. Platform (recurring, predictable)
The easiest part to estimate and the least decisive.
| Platform | Billing model |
|---|---|
| Shopify | Fixed monthly plan, plus a transaction fee outside its own gateway |
| VTEX | Percentage of revenue (take rate), with a monthly floor |
| Magento (open source) | No license; you pay for infrastructure and maintenance |
| WooCommerce | No license; you pay for hosting, plugins and maintenance |
The common mistake is choosing by license cost. Magento and WooCommerce look free and are not: what you save in licensing, you spend on infrastructure and on whoever keeps it running.
2. Implementation (one-off, variable)
This is where proposals differ most. What moves the number inside the R$ 25k–45k range:
- Catalog complexity. A thousand simple SKUs cost less than two hundred with variants, kits and pricing rules.
- Number of capabilities. Loyalty, B2B, subscription and marketplace are scope units, each with its own implementation on the chosen platform.
- Storefront customization. An adjusted kit is one thing; bespoke design is another — and it belongs to the design line, not this one.
- Migration. Bringing catalog, customers, orders and URLs from an existing store costs more than people expect.
3. Integrations (where budgets break)
The most underestimated line, and the one that must appear separately in the proposal. A store that does not talk to the ERP turns into manual order entry within three weeks.
The usual set:
- ERP — product, price, stock, orders and invoicing;
- Shipping — quoting and tracking, sometimes across carriers;
- Payments — gateway, anti-fraud, installments;
- Invoicing — automatic issuance after payment;
- Marketing — GA4, media tags, email and messaging.
Not all of them need to exist on day one. All of them need to be in the plan, because fitting an integration in after launch costs more.
4. Ongoing work (recurring, almost always forgotten)
After launch the store does not stand still. Seasonal campaigns, shipping rule changes, a new product with a different structure, platform updates, a bug that only shows up at volume.
That R$ 3,000 to 5,500 a month buys people's time, not hosting. It is the most common wrong comparison: someone puts this line next to "a server for R$ 500" and concludes it is expensive — but one is a machine and the other is whoever answers when orders stop reaching the ERP on a Friday night.
A project priced without this line produces, three months later, an unpleasant conversation about "additional scope" — or a store frozen in the state it was born in.
What about the timeline?
Timelines depend less on the vendor and more on two things: how many integrations exist and how ready your content is. The ranges we see:
| Scope | Typical timeline |
|---|---|
| New store, catalog ready, no integrations | 4 to 6 weeks |
| Store with ERP, invoicing and shipping integrated | 8 to 12 weeks |
| Platform migration with a large catalog and order history | 3 to 5 months |
Delays almost never come from development. They come from:
- catalog with no photos or descriptions;
- ERP access that takes weeks to be granted;
- a shipping decision nobody wants to make;
- invoicing homologation.
The part that depends on code is the most predictable part of the project. The part that depends on data and decisions is what slips. If you want a shorter timeline, the answer is not pressuring the vendor — it is having photos, descriptions and ERP access ready before the project starts.
Where projects spend badly
Design before operations. Investing heavily in a unique storefront before solving stock and shipping is the most expensive and most common mistake.
Half-built integrations. A connector that pulls products but never pushes orders creates permanent manual work.
Platform above your size. Buying an enterprise solution for a three-hundred-item catalog pays for complexity and receives nothing back.
No measurement. A store without GA4 configured properly is a store that does not know where it loses sales — and every decision after that is a guess.
How to evaluate a proposal
Ask for the four lines separately. If implementation, integration and ongoing work arrive as a single number, you cannot compare it, negotiate it, or understand what happens when scope changes.
And ask what happens the month after go-live. That answer tells you more about the vendor than the total does.
Frequently asked questions
How much does an e-commerce project cost?
For a company that already sells and has no development team, implementation runs between R$ 25,000 and R$ 45,000. Adding design and branding, a typical full project lands around R$ 40,000, before complex ERP integration work.
How much does it cost to run an online store per month?
Two lines add up: the platform (a Shopify plan, a VTEX take rate, or infrastructure in the case of Magento and WooCommerce) and technical retainer work, which runs from R$ 3,000 to R$ 5,500 a month for mid-market operations.
How long does it take to launch?
Four to six weeks when the catalog is ready and there are no integrations. Eight to twelve weeks with ERP, invoicing and shipping. Three to five months for a migration with a large catalog and order history.
Why does one proposal cost R$ 8k and another R$ 40k?
Almost always because one of them excludes integration, data migration or post-launch support. Compare line by line; if the quote arrives as a single number, there is nothing to compare.
Can we launch without ERP integration?
Yes, and sometimes that is the right way to validate the operation. But keep the integration in the plan from the start: fitting it in later costs more than doing it alongside, and the manual work in the meantime has a cost of its own.
Want a number for your case? Tell us how many SKUs you have, which systems you use and where you sell today. Castle replies with a concrete path, not a generic quote. Talk to us.