Custom E-commerce vs. Shopify: Which Is Right for Your Indian Business?

This isn't a "Shopify is bad, custom is good" post — Shopify is a genuinely solid platform for a huge range of stores, and for plenty of businesses it's the right call. The honest question isn't which one is better in the abstract; it's which one fits where your store actually is, and where it's headed.

What Shopify gets right

Shopify is fast to launch, has a mature app ecosystem for almost anything a standard store needs, and doesn't require an engineering team to run day to day. For a store just getting started, with no dedicated technical resource, that's a real advantage — you can be selling within days, not weeks.

Where the trade-offs start to bite

Recurring costs stack up

The monthly plan fee is only the starting point. Apps for AI recommendations, advanced search, abandoned-cart flows beyond the basics, and A/B testing each carry their own subscription — and if you're not on Shopify Payments, a transaction fee applies on top of your payment gateway's cut. A store running six or seven apps to cover functionality a custom build would ship natively can easily be paying more per month, indefinitely, than a one-time custom build would have cost outright.

Theme and app-marketplace ceilings

Need a checkout flow that doesn't match Shopify's constraints? A product configurator that doesn't exist as an app? A recommendation engine tuned to your specific catalog logic rather than generic co-purchase data? Each of those either isn't possible on standard Shopify plans, or means paying for Shopify Plus and a development partner anyway — at which point you're paying custom-build prices without getting a custom-build result.

You don't own the platform

Your store lives inside Shopify's infrastructure, under Shopify's pricing decisions and policy changes. That's a reasonable trade for the convenience it buys — but it is a trade, and it's worth being clear-eyed about it rather than discovering it later.

What a custom-coded storefront costs instead

• Starter — ₹25,000: an MVP Next.js storefront, up to 25 products, Razorpay or Stripe Checkout, basic AI-powered search, 14-day delivery
• Growth — ₹75,000: a fully custom storefront with a lightweight admin panel, up to 250 products, AI product recommendations, abandoned-cart recovery, multi-gateway payments, A/B testing, and basic email automation — 30-day delivery
• Pro AI — ₹1,99,999: full admin dashboard, up to 2,500 products, an advanced AI personalization engine, multi-vendor capability, marketing automation, and an analytics dashboard — 60-day delivery
• Enterprise — from ₹7,50,000: headless architecture, custom AI/ML models trained on your data, multi-currency and multi-region, SLA and a dedicated team

No transaction fee stacked on top of the payment gateway. No app subscriptions replacing what's built in natively. A higher one-time cost, and no recurring platform bill for the core functionality.

A practical way to decide

Shopify is the right call if you're validating a new product line, have no dedicated technical resource, and need to be live in days rather than weeks. A custom build starts making more sense once you're paying for four or more apps just to get functionality you actually need, once your catalog or checkout logic doesn't fit standard theme patterns, or once the monthly platform-plus-apps bill starts approaching what a one-time custom build would have cost.

Migrating from Shopify to a custom build

This doesn't have to be an all-or-nothing decision made once and never revisited. Product data, order history, and customer records can typically be exported from Shopify and migrated into a custom catalog when the time comes — the scope depends on catalog size and how much custom logic is involved. Plenty of stores start on Shopify to validate demand, then move to a custom build once the platform's ceiling starts costing more than a rebuild would.

What changed with international expansion

The tier prices above now have USD equivalents too — $750 Starter, $2,250 Growth, $6,000 Pro AI, from $22,500 for Enterprise — for stores outside India. Razorpay stays the default for Indian transactions; Stripe handles the international side, running in parallel rather than replacing one with the other. A store serving both markets doesn't need two separate builds, just both gateways wired into the same checkout.

The AI features gap, concretely

This is where the platform-versus-custom gap gets most visible. Shopify's AI recommendation apps mostly work off generic co-purchase data — what other stores' customers bought together, not what your specific catalog and margins actually support. A custom build's recommendation engine is tuned to your catalog from day one, with no separate app subscription stacking on top of the platform fee. Same story with search: a generic Shopify search app matches keywords; a custom build can run vector search with LLM re-ranking that understands a query like "something for a rainy day gift under $60" well enough to return the right products, not just ones with matching text.

What the first 90 days actually look like

Starter and Growth both ship with post-launch support built in — 30 days on Starter, 60 on Growth — covering bug fixes and the inevitable small adjustments once real customers start using the store. That window matters more for e-commerce than for a brochure site, since checkout and payment issues surface under real transaction volume in a way a design review never catches. Pro AI extends that to 90 days, given the larger scope has more surface area to settle.

A note on catalog size, honestly

None of the tiers above are arbitrary product-count ceilings picked to upsell. 25 products on Starter is genuinely enough for a lot of first e-commerce builds — a founder testing one product line doesn't need a 2,500-product architecture sitting unused. The jump to Growth's 250 products usually tracks with a store that's validated demand and is actively expanding the catalog, not a number chosen to force an upgrade. A catalog that genuinely needs more than 2,500 products, or multi-vendor architecture, moves to Enterprise, scoped individually rather than forced into a tier that wasn't built for it.

Why the discovery call still matters

Picking a tier off this page alone skips the part that actually prevents mismatches: a real conversation. The free 30-minute call exists to catch them early — a catalog that looks like Starter on paper but actually needs Growth's admin dashboard because three people will be managing inventory, or a Growth-sized store that could actually ship on Starter because the "AI recommendations" ask turns out to be a nice-to-have, not a launch requirement. That conversation upfront saves a scope conversation happening after a quote's already been signed.