Key takeaways
- Compare five years of ownership, not the launch quote.
- A custom build moves cost from fees to engineering, security and maintenance.
- A platform reduces undifferentiated work but limits how far you can bend it.
- Your real differentiator should decide where you customize.
You are choosing how to run online sales. One option is a managed commerce platform such as VTEX, with recurring fees and a defined feature set. The other is a custom build, which promises total control. The launch quotes look very different, and that comparison is misleading. What matters is what each path costs to own over several years. We deliberately use no invented numbers here. Use your own quotes and this structure.
Where the money actually goes
A platform charges you for the service: typically a subscription or a share of sales, plus implementation, integrations and support. A custom build charges you for people: developers, designers, quality assurance, security, hosting and on-call time. The first cost is visible and recurring. The second is often underestimated because it continues after launch.
Cost lines to compare side by side
- Initial build or implementation, including design, integrations and data migration.
- Recurring platform fees or hosting and infrastructure.
- Payment, tax, shipping and search integrations, and their upkeep.
- Security, compliance and payment-standard obligations.
- Upgrades: platform releases versus framework and dependency updates you own.
- Performance work for peak traffic, including load testing.
- Internal team time for product, operations and support.
- Cost of delay: what a slower launch means for sales.
What a managed platform typically gives you
Checkout, catalog management, promotions, order management and many marketplace and payment connectors are already built and maintained by the vendor. That shortens the time to launch and shifts patching and scaling to someone else. The trade-off is that you work within the platform's model. Unusual pricing rules, special workflows or unconventional experiences may need workarounds or extensions.
What a custom build typically gives you
Freedom to design exactly the experience and business logic you want, and no dependence on one vendor's roadmap. The trade-off is that you own everything: bugs, security patches, scaling, integrations and the knowledge held by a small number of engineers. If those people leave, cost and risk rise quickly.
Questions that break the tie
- What is our real differentiator online: the catalog, the price logic, the experience, or the service? Customize there and buy the rest.
- How many engineers can we dedicate, and for how many years?
- How fast must we launch, and what does each month of delay cost?
- How many channels and marketplaces do we need to connect?
- What happens to the business if a key developer leaves?
A middle path
Many teams pick a platform for the commodity core and build custom pieces around it, such as a tailored storefront, pricing service or loyalty logic. This keeps the risky parts, such as checkout and payments, on maintained ground while still expressing what makes the business different.
How to run the comparison
Build a simple five-year model with the cost lines above for each path. Add a low, expected and high scenario for traffic and team turnover. Ask vendors and agencies which items are excluded from their quotes. The winner is often clear once maintenance and delay costs are on the page.
