Building commerce when the platforms genuinely do not fit
Custom eCommerce means building the store rather than configuring a platform. It is justified when your product model, pricing logic or fulfilment process cannot be expressed in what Shopify or WooCommerce provide. And it is the wrong answer when they can.
Most businesses asking for custom eCommerce should use a platform, and the useful part of this conversation is establishing which group you are in. The platforms have absorbed a decade of edge cases. Deciding to rebuild what they give you free needs a reason more specific than wanting control.
The genuine cases share a shape. Pricing that depends on customer, quantity, contract and configuration simultaneously. Products configured rather than chosen, where the configuration determines price and lead time. Inventory across several locations with allocation rules. Fulfilment split across suppliers with different terms. These are expressible in a platform only through increasingly elaborate workarounds, and the workarounds eventually cost more than the build.
When this work makes sense
Questions worth asking any supplier
Has anyone properly tested whether a platform can do this? That evaluation saves the most money.
Is the complexity in pricing, product model or fulfilment? Each has different platform workarounds worth trying first.
What are you giving up? Platforms provide fraud handling, PCI scope reduction and checkout optimisation you will now own.
Who maintains it in three years? Custom commerce needs an owner permanently.
Could the complex part be a service alongside a platform storefront? Often the cheapest answer.
Deliverables
A written build-versus-buy assessment including the case for using a platform
Data model for products, variants, pricing rules and inventory
Storefront and checkout built around your actual purchase flow
Pricing engine expressing your real rules rather than approximating them
Payment integration with PCI scope deliberately minimised
Inventory and fulfilment logic matching how you operate
Admin tooling so your team manages the store without developers
Order, refund and exception handling designed rather than discovered
Built with
Process
The first deliverable is an assessment of whether a platform could do this, written down and given to you, including the case against building. We test the requirement against Shopify and WooCommerce specifically and document where each falls short. Losing the project to that document is a better outcome than building a custom store for a business that needed a configured one.
The standard nine stages, discovery, strategy, UX/UI, development, content and SEO, testing, launch, measurement and continuous improvement, apply to every project. The paragraph above is what differs for this one.
What moves the number
Pricing depends on scope, functionality, integrations and content requirements. These are the factors that change it most:
Platform, platform plus custom services, or fully custom
Platform as-is
Fits: The large majority of retail businesses
Limits: Your model must fit what the platform expresses
Platform plus custom services
Fits: Businesses with one genuinely unusual area such as pricing, configuration or allocation
Limits: Integration complexity between the platform and your service
Fully custom
Fits: Businesses whose entire commerce model is unusual
Limits: You rebuild fraud handling, checkout optimisation and PCI work the platforms provide
Case studies
No published custom commerce case study. These builds are usually commercially sensitive to the businesses running them.
MBGDesk operates from 4 offices in India, United States, United Kingdom, United Arab Emirates. We hold no industry certifications and do not imply otherwise.
Frequently asked questions
How do we know if we need custom eCommerce?
Try to express your requirement in a platform first, properly rather than as a thought experiment. If it needs three apps and a workaround that someone has to remember, custom may be justified. If it works with configuration, use the platform.
What about headless commerce?
Often the middle answer. You keep a platform's commerce engine, catalogue, checkout, payments, and build a custom front end. You get presentation freedom without rebuilding the parts that are genuinely hard and genuinely solved.
Is custom more expensive?
To build, substantially. Over five years it depends on what platform fees and workaround maintenance cost you. For most businesses the platform still wins; for a few with real complexity, the workarounds cost more than the build would have.
What about PCI compliance?
Use a payment provider's hosted fields or redirect so card data never touches your servers. That keeps your PCI scope small. Handling card data directly is a compliance burden almost no business should take on deliberately.
Can you build B2B features?
Yes, account-specific catalogues and pricing, quantity breaks, credit terms, purchase order workflow, approval chains. This is where platforms most often fall short and where custom is most frequently justified.
What happens if we outgrow it?
You extend it, which is the argument for custom. The failure mode is different from platforms: you hit no ceiling, you accumulate technical debt. Whether that is better depends on whether someone maintains it.
Could we start on a platform and move later?
Yes, and many should. Prove the business on a platform, migrate when the workarounds genuinely cost more than a build. Migration is real work, and it is a smaller risk than building custom for a model you have not validated.
How do you handle search on a custom store?
With a dedicated search index rather than database queries, which is the part most custom builds underestimate. Shoppers expect typo tolerance, synonyms, faceted filtering and relevance ranking, all of which a platform provides and a custom build must implement. It is a meaningful piece of work in its own right, and store search that returns nothing for a misspelled product name costs sales quietly every day.
Services that connect to this
Talk to us about this
Tell us what you are trying to build and we will come back with scope, approach and an estimate.