You have a product ready. Maybe 300 units sitting in your spare room, maybe a supplier waiting on your first purchase order. One question is blocking everything: do you list on Amazon, or build your own Shopify store first? Everyone around you has an opinion. The Amazon camp says the traffic is already there. The Shopify camp says marketplaces eat your margin and keep your customer. Both are partly right, which is exactly why the debate never ends on WhatsApp groups.
This guide settles it with numbers instead of opinions. We will put the 2026 fee structures side by side, run the full margin math on the same ₹599 product on both platforms, compare who owns the customer and the data, look honestly at COD and RTO exposure on each, and end with a clear decision framework: when Amazon first, when Shopify first, and when both.
If you are still one step earlier, deciding what to sell rather than where, read how to start an ecommerce business in India first and come back. This guide assumes you have a product and roughly ₹30,000 to ₹1,00,000 to deploy.
For most first-time Indian sellers in 2026, Amazon is the cheaper place to win your first few hundred orders. Referral fees dropped to zero on products under ₹1,000 across 1,800+ categories from March 2026, so a ₹599 product can net around ₹200 per order on FBA. Shopify costs more per new customer, roughly ₹110 net on the same product once CAC and RTO are counted, but every buyer, their phone number and every repeat order belongs to you, so a repeat order nets ₹270 or more. The honest answer: validate demand where the traffic already is, build the brand where the customer is yours, and run both once you cross roughly 10 orders a day.
What you are actually choosing between
A marketplace like Amazon is a shop inside someone else's mall: the footfall, the checkout, the delivery network and the customer relationship are theirs, and you pay a slice of every sale for access. An own store on Shopify is your own shop on your own plot: you pay fixed rent and bring every visitor yourself, but everything that happens inside, the data, the margin structure, the repeat sale, is yours.
Note that this is a distribution decision, not a business model decision. Whether you hold inventory or not is a separate question; if you are still weighing that, read dropshipping vs D2C in India before picking a channel.
Amazon vs Shopify fee structure in India (2026)
Start with what each platform actually charges, because 2026 changed the math meaningfully on the Amazon side.
What Amazon charges sellers in 2026
Amazon India charges three core fees plus 18% GST on all of them: a referral fee (a percentage of the sale price, set per category), a closing fee (a flat per-item charge by price band), and a fulfilment fee (weight handling if Amazon ships it via FBA or Easy Ship). From March 16, 2026, Amazon expanded zero referral fees to products priced under ₹1,000 across more than 1,800 categories, covering roughly 12.5 crore listings in apparel, beauty, home, kitchen, toys, grocery and more. Closing fees also fell: under ₹300 the fee dropped from ₹45 to ₹20, and in the ₹300 to ₹500 band from ₹35 to ₹26, with Easy Ship fees cut by over 20% on products under ₹300, per independent breakdowns of the 2026 fee update. Above ₹1,000, referral fees still apply, typically anywhere from around 2% to 16% depending on category, so always check the live rate card on Amazon's fees and pricing page for your exact category.
What Shopify charges in India in 2026
Shopify charges a fixed subscription instead of a commission. Per Shopify's India pricing page, the Basic plan is ₹1,994 per month on monthly billing (₹1,499 per month if you pay yearly), Grow is ₹7,447 and Advanced is ₹30,164 on monthly billing, all plus 18% GST. The same page lists a transaction fee on third-party payment providers, 2% on Basic, on top of whatever your gateway charges. Since you collect payments through an Indian gateway like Razorpay, Cashfree or PayU, budget roughly 2% plus GST on the fee for prepaid orders. Shipping is on you too: courier aggregators price by 500 gram slab and zone, so a light parcel costs anywhere from about ₹35 within your region to ₹90 or more across the country.
| Fee head | Amazon (seller) | Shopify (own store) |
|---|---|---|
| Fixed monthly cost | ₹0 on the basic selling plan | ₹1,994 + GST (Basic, monthly billing) |
| Commission on sale | 0% under ₹1,000 in 1,800+ categories; roughly 2% to 16% otherwise | None |
| Closing fee | ₹20 to ₹45 flat per item by price band | None |
| Payment collection | Included in fees | Gateway ~2% + GST, plus 2% Shopify transaction fee on Basic |
| Fulfilment and shipping | FBA or Easy Ship weight handling, roughly ₹50 to ₹100 for a 500g parcel | Courier aggregator, roughly ₹35 to ₹90 per 500g by zone |
| Storage | FBA monthly storage per cubic foot | Your own space, often ₹0 early on |
| Marketing | Amazon Ads (optional but usually needed) | Meta and Google ads (almost always needed) |
| GST on platform fees | 18% on all fees | 18% on subscription and gateway fees |
Margin per order on a ₹599 product: the honest math
Fee tables hide the truth; a worked example exposes it. According to the Margin Waterfall™ framework, you calculate what survives of the selling price after every real cost, including returns and marketing, before you commit a single rupee of ad budget. Let us run the same product through both platforms: a ₹599 home and kitchen item, landed COGS of ₹180, parcel weight under 500 grams, sitting in a category eligible for Amazon's zero referral fee.
Margin Waterfall™: selling price minus COGS, packaging, shipping, payment or platform fees, RTO loss, then CAC. If the number at the bottom is negative, no amount of scale saves it.
Assumptions for the Shopify column: Basic plan on monthly billing amortised over 150 orders a month, gateway at 2% plus GST on the fee, blended courier cost of ₹70, a first-order CAC of ₹150 on Meta ads (a realistic blended figure for a new brand with decent creative), and an RTO provision explained in the RTO section below. For Amazon: FBA fulfilment around ₹77 for this weight, closing fee of roughly ₹35 in the ₹500 to ₹1,000 band, amortised storage, and about ₹60 of Amazon Ads per order, since page one placement rarely comes free even with zero referral fees.
| Line item | Amazon FBA | Shopify (own store) |
|---|---|---|
| Selling price | ₹599 | ₹599 |
| Referral fee | ₹0 (under ₹1,000, eligible category) | Not applicable |
| Closing fee | ~₹35 | Not applicable |
| Fulfilment / courier (≤500g) | ~₹77 (FBA) | ~₹70 (aggregator, blended) |
| Storage, amortised | ~₹8 | ₹0 (you store it) |
| Subscription, amortised over 150 orders | ₹0 | ~₹16 |
| Payment gateway (2% + GST on fee) | Included | ~₹14 |
| Shopify transaction fee (2%, Basic) | Not applicable | ~₹12 |
| GST 18% on platform fees | ~₹22 | Included above |
| Packaging | ~₹10 (inner only) | ~₹25 (full outer) |
| COGS | ₹180 | ₹180 |
| Marketing per order | ~₹60 (Amazon Ads) | ~₹150 (Meta CAC) |
| RTO / returns provision | ~₹5 | ~₹20 |
| Net per first order | ~₹202 | ~₹112 |
Three honest caveats before you screenshot this table. First, the Amazon number depends heavily on the zero referral fee: if your category is not eligible, a typical 13% referral fee plus GST removes about ₹92 and drops the Amazon net to roughly ₹110, near parity with Shopify. Second, the Shopify CAC of ₹150 is earned, not given; weak creative or a crowded category can push it to ₹250 and wipe the margin out entirely. Third, both columns move with your inputs, which is why you should model your own product with the full method in D2C unit economics in India before ordering inventory.
Now the part the first-order table hides: the repeat order. On Shopify, a repeat customer arrives free via WhatsApp or email, usually prepays, and rarely RTOs. That order nets around ₹272 on the same product. On Amazon, the repeat order looks identical to the first one, still roughly ₹202, because the customer belongs to Amazon and your competitor's sponsored listing sits one thumb-scroll above your product. Amazon wins the first sale; Shopify compounds.
Who owns the customer: the gap nobody prices in
The fee tables above compare what each platform takes. This table compares what each platform gives back, and it is where the two roads truly separate.
| Question | Amazon | Your Shopify store |
|---|---|---|
| Do you get the buyer's phone and email? | No, Amazon masks them | Yes, on every order |
| Can you retarget past buyers with ads? | Only inside Amazon's ad system | Yes, Meta and Google pixels are yours |
| Who decides if the repeat order is yours? | Amazon's search results and buy box | Your WhatsApp list and email flows |
| Pricing control | Constant pressure from fee changes and competing offers | Full control, bundles and memberships included |
| Platform risk | Account suspension can freeze the whole business overnight | Low; gateway or domain issues are rare and recoverable |
| Reviews and social proof | Locked to the listing, not portable | Portable across your site, ads and Instagram |
This is why experienced operators describe Amazon revenue as rented and website revenue as owned. A brand doing ₹5 lakh a month purely on Amazon has a P&L; a brand doing ₹5 lakh a month on its own store has a P&L plus an asset: a customer list that lowers every future month's acquisition cost.
COD and RTO exposure: Amazon vs your own store
RTO (return to origin) is an order that ships but never gets delivered, refused at the door or unclaimed, and comes back at your cost. In India this is mostly a COD problem, and the exposure differs sharply by channel. On your own store, industry data is blunt: Shipway's ShipNotes report found a 26% RTO rate on COD orders across India, against low single digits for prepaid, and COD-heavy categories like fashion can touch 40%.
On Amazon, COD refusal still exists but hits you softer: Amazon's delivery network, doorstep trust and pre-delivery communication keep undelivered rates low, and with FBA the reverse leg is handled inside the system, though you still bear return processing and the occasional damaged unit. On your own store, every RTO costs you both courier legs, the packaging and the CAC you already spent, easily ₹150 to ₹250 per failed order. That is why the Shopify column above carries a ₹20 per-order provision, assuming roughly 40% of orders on COD and a quarter of those returning; run your own mix, and read how to reduce RTO on COD orders before you turn COD on at all.
Discoverability vs brand building
Amazon is demand capture: crores of buyers already search generic terms like "steel water bottle 1 litre" with a card saved and an address filled. You do not create the demand, you intercept it, which is why a good listing can sell from week one. The cost of that convenience is anonymity: you are a listing among thousands, one fee change or one aggressive competitor away from losing the shelf.
Shopify is demand creation: nobody wakes up and types your domain name in month one. You earn every visitor through Meta ads, Google, Instagram content and word of mouth. It is slower and more expensive per order at the start, but every rupee also builds something Amazon never gives you: recall, a list, a story people buy into at full price. If you go this route, your ad system is the business; start with Meta ads for D2C in India and treat CAC as your most-watched number.
As Ravikant Tyagi, who ran distribution at Eureka Forbes and supply chain and operations at Atomberg before working with D2C founders as a fractional COO, puts it:
I have watched founders burn three months and a lakh of rupees building the perfect website for a product nobody had validated, while a competitor listed a rough version on Amazon, read 200 reviews, fixed the product, and then built the website. The marketplace is not your enemy in year one. It is the cheapest market research machine in India, paid for by orders instead of surveys.
Launching Shopify first with an unvalidated product and treating Meta ads as a demand switch. A typical first-timer spends ₹15,000 on the store setup and ₹40,000 on ads at a ₹250+ CAC because the creative and the offer are untested, nets under ₹50 an order before RTO, and concludes "D2C does not work." The product was never validated; the founder paid Meta to find that out at roughly ₹55,000 when 40 Amazon orders could have said the same thing while making money.
The decision framework: Amazon first, Shopify first, or both?
According to the Founder Decision Loop™, channel selection comes after demand proof, because a perfect store for a product nobody wants is still a loss. The loop runs decide, deploy small, measure, correct, and the channel you pick first is simply the one that gives you the cheapest, fastest read on real demand for your specific product.
Founder Decision Loop™: frame the decision, commit a small reversible bet, set the number that proves or kills it, and review on a fixed date. Applied here: pick one channel, 100 units, 45 days, and a target margin per delivered order. The data, not the debate, picks your second channel.
| Your situation | Start here | Why |
|---|---|---|
| Product under ₹1,000 in a zero-referral category, no audience | Amazon first | Zero commission plus built-in traffic makes it the cheapest validation lab in 2026 |
| Generic, search-driven product (bottles, organisers, cables) | Amazon first | Demand already exists as search volume; intercept it |
| COD-heavy impulse category and you fear RTO | Amazon first | Amazon's delivery trust absorbs most of the COD refusal risk you would carry alone |
| Story-led, premium or personalised product, AOV ₹1,500+ | Shopify first | Referral fees bite above ₹1,000 and the brand story cannot breathe inside a listing |
| You already have 10,000+ engaged Instagram followers | Shopify first | You own distribution already; do not pay a mall for footfall you can bring yourself |
| Repeat-purchase category (skincare, supplements, coffee) | Shopify first, Amazon as shelf | LTV lives in the owned list; the marketplace is discovery, not the relationship |
| Stable 10+ orders a day on one channel | Add the other | You have proof; now Amazon adds reach while Shopify adds ownership |
If your product sells under ₹1,000 in an eligible category and you have no audience → list on Amazon first, 100 units, 45 days. If your product is premium, story-led or repeat-purchase and you can fund a ₹150 to ₹250 CAC → build Shopify first with a prepaid-first setup. If Amazon gives you 10 orders a day at positive margin → launch Shopify to convert rented demand into an owned list. If neither channel produces a positive Margin Waterfall™ on paper → fix price or product, not the platform.
How much money each path actually needs
Amazon first: expect roughly ₹25,000 to ₹60,000 to start, mostly inventory, GST registration, product photos and a small Amazon Ads budget; there is no monthly platform rent on the basic selling plan. Shopify first: expect roughly ₹60,000 to ₹1,20,000 for the same product, since you add the subscription, a theme or apps, packaging that carries the brand, and a real ad budget, because the store without traffic is a brochure. The full breakdown by line item is in how much it costs to start a D2C brand in India.
- Check whether your category and price point qualify for Amazon's zero referral fee on the live rate card
- Run the Margin Waterfall™ for your product on BOTH channels before choosing
- Get GST registration done; both paths need it
- Amazon path: professional photos, keyword-rich listing, FBA for your first batch
- Shopify path: gateway approval (Razorpay or Cashfree), courier aggregator account, WhatsApp order confirmations on from day one
- Set a 45-day validation window with a written kill number (minimum net margin per delivered order)
- Cap COD exposure early: prepaid discount on your own store, COD verification before dispatch
- Whichever channel wins, start collecting customer phone numbers legally from day one
Your next action
Today, open two tabs: Amazon's fee rate card for your category and Shopify's India pricing page. Fill one sheet with your product's Margin Waterfall™ on each channel, using your real COGS and honest CAC and RTO assumptions, not hopeful ones. The channel with the higher net margin per delivered order at your current stage is your first channel. That is one hour of work, and it replaces months of second-guessing.
If you'd like the complete execution system, calculators, SOPs, templates and operating frameworks behind this process, continue inside D2C Acquisition.Lab.
