You have decided to sell online, and now you are stuck in the gap between the decision and the first order. Do you need a company before you can sell? Does GST apply from day one or only after ₹40 lakh? Should you list on Amazon or build your own store? Every article online answers one slice, half of them contradict each other, and meanwhile your product idea sits in a notebook.
This guide is the complete A to Z roadmap for starting an ecommerce business in India in 2026, in the exact order an operator would sequence it: demand proof first, legal structure second, unit economics before inventory, platform choice before store design, and paid marketing last. Every step carries a realistic ₹ cost and a timeline, so by the end you can write down your own launch budget instead of guessing.
It is written from the operator's side of the table by Ravikant Tyagi, who has run national distribution at Eureka Forbes, led supply chain and operations at Atomberg, and now works as a fractional COO for Indian D2C brands. The eight steps below follow the same sequence used inside our execution system, compressed into one guide.
You can start an ecommerce business in India with ₹40,000 to ₹1,50,000 and be live in 30 to 45 days. The sequence that works: validate demand with ₹2,000 to ₹5,000 before building anything, register as a sole proprietor and take GST only when the rules actually require it, lock per-order profit math before paying a supplier, choose marketplace or own store based on your price point and repeat-purchase potential, connect Razorpay for UPI and cards, ship through a courier aggregator at roughly ₹36 to ₹45 blended per 500g order, and spend your first ₹10,000 of ads learning your acquisition cost, not chasing revenue.
Why 2026 is a genuinely good year to start selling online in India
The macro tailwind is real. Indian ecommerce is projected to reach around US$ 163 billion by 2026, growing at roughly 27 percent a year, according to IBEF. UPI now settles the majority of online payments by volume, logistics aggregators have pushed nationwide shipping under ₹50 for a light parcel, and both Amazon and Flipkart have dropped commissions to zero on large ranges of products under ₹1,000 to attract small sellers.
None of that guarantees your survival. The same low barriers that let you launch in a weekend let ten competitors do the same. What separates the stores that last from the ones that die by month six is not the product or the logo. It is sequencing and math, which is exactly what this roadmap enforces.
How much does it cost to start an ecommerce business in India?
Before the steps, the honest budget. These are 2026 numbers for a lean, single-founder launch with a physical product.
| Item | Realistic cost | When you pay |
|---|---|---|
| Demand validation (landing page + small ad test) | ₹2,000 to ₹5,000 | Week 1 to 2 |
| Registration (proprietorship + GST via a CA) | ₹0 to ₹5,000 | Week 2 |
| Samples + first inventory (50 to 150 units) | ₹15,000 to ₹60,000 | Week 3 to 4 |
| Packaging (boxes, filler, tape, inserts) | ₹3,000 to ₹8,000 | Week 3 to 4 |
| Store: Shopify Basic + domain | ₹2,300 first month | Week 4 |
| Product photos (phone + daylight + a friend) | ₹0 to ₹5,000 | Week 4 |
| Marketing test budget | ₹10,000 to ₹25,000 | Week 5 to 6 |
| Trademark, class 35 or your product class (optional now) | ₹4,500 govt fee | Any time |
Total: roughly ₹40,000 at the frugal end, ₹1,50,000 if you order more inventory and test ads properly. The trademark figure is the official fee for individuals and MSMEs per IP India; it can wait until revenue exists. For a line-item breakdown of where each rupee goes, see our guide on the real cost to start a D2C brand in India.
Step 1: Validate demand before you build anything (Week 1 to 2 · ₹2,000 to ₹5,000)
Demand validation is the act of getting strangers to show buying intent for your product before you own inventory or a store. It is the cheapest insurance in ecommerce, and it is the step most founders skip because building a store feels like progress and asking strangers for money feels uncomfortable.
According to the Founder Decision Loop™, demand validation comes before supplier selection, because a great supplier for a product nobody wants is still a loss. The loop forces every launch decision through the same gate: evidence first, spend second.
Founder Decision Loop™: Observe (what are buyers already paying for, and complaining about), Commit small (the cheapest test that produces real intent, a ₹2,000 ad to a waitlist page, a pre-order to 20 people), Measure (did strangers click, pre-pay, or share a phone number), then Decide (scale the bet, fix the offer, or kill it). No step in this roadmap gets funded until the previous loop closes with evidence.
Practical validation tests that cost less than one month of Shopify:
- Marketplace autopsy. Study the top 10 listings for your product on Amazon and Flipkart. Review velocity, price clusters, and one-star complaints tell you demand, price ceiling, and your opening angle, in that order.
- A one-page test. A free landing page with your product, price, and a "Pre-order" or "WhatsApp to order" button, plus ₹1,500 to ₹3,000 of Instagram traffic. You are measuring clicks to the buy action, not compliments.
- 20 real conversations. Not friends. People who currently buy the competing product. Ask what they paid, what annoys them, and whether they would switch at your price.
The full method, including pass and fail thresholds, is in our guide on how to validate a business idea before spending money.
Building the store before validating demand. A founder spends three weeks on themes, logo revisions and packaging mockups, then orders 500 units on instinct, and only then finds out the market wants the product at ₹399 while their landed cost needs a ₹649 price. The store looked ready; the business never was. Typical damage: ₹40,000 to ₹80,000 locked in dead stock plus two months of morale. The store is a two-day job in Step 5. Demand proof is the hard part, which is exactly why it comes first.
Step 2: Choose your legal structure and register (Week 2 · ₹0 to ₹15,000)
You do not need a private limited company to start selling online in India. You need a legal identity, a current account, and the right tax registrations for how you sell. Here is the honest comparison for a first-time founder in 2026, with all-in costs including professional fees, per RegisterKaro's 2026 fee breakdown.
| Structure | All-in setup cost | Best for | Watch out for |
|---|---|---|---|
| Sole proprietorship | ₹0 to ₹5,000 | Almost every first launch | No liability shield, harder to raise funds |
| LLP | ₹8,000 to ₹15,000 | Two co-founders, services + products | Annual filings even at zero revenue |
| Private limited | ₹7,000 to ₹25,000 | Raising money, ESOPs, brand contracts | Compliance cost of ₹15,000+ every year |
Start as a proprietorship unless you have co-founders or an investor conversation already running. You can convert later; thousands of brands have.
When do you actually need GST registration?
This is the most misunderstood part of ecommerce registration in India, so here is the 2026 position in plain language, consistent with the thresholds explained by Tally Solutions:
- Selling on Amazon, Flipkart or Meesho: treat GSTIN as mandatory. Marketplaces deduct 1 percent TCS on your sales and verify GSTIN at onboarding for most categories. A 2023 relaxation lets very small intra-state sellers list without GST on some platforms, but the big marketplaces enforce GSTIN for the vast majority of sellers in practice.
- Your own website, shipping across states: inter-state supply of goods generally makes registration mandatory regardless of turnover. Since almost every online store ships pan-India, assume you need it.
- Your own website, delivering only within your state: the standard threshold applies, ₹40 lakh annual turnover for goods in most states, ₹20 lakh in special category states.
Registration itself is free on the GST portal; a CA typically charges ₹1,500 to ₹3,000 to do it cleanly. Confirm your exact case with a CA, because category exceptions exist.
Before you apply: keep PAN, Aadhaar, a photo, a cancelled cheque of the account you will actually use for the business, and proof of your principal place of business (electricity bill plus rent agreement or NOC if the property is not yours). Apply as "supplier of goods through e-commerce operator" where applicable, pick your state carefully because it decides your CGST/SGST jurisdiction, and expect approval in 3 to 7 working days if documents match. Mismatched name spellings between PAN and the bank account cause most rejections.
Step 3: Find a supplier and lock your unit economics (Week 2 to 4 · your biggest spend)
With demand evidence in hand, source the product. For a first run, your options in rough order of control: buy wholesale from markets like Sadar Bazaar or your city's equivalent, source small MOQ private label via IndiaMART or direct manufacturer visits, or white-label an existing product with your branding. Get samples from at least three suppliers, pay for them, and test them the way a rough courier journey would.
Two rules protect you here. First, the 3x rule: if you cannot sell at roughly three times your landed cost (product plus packaging plus inbound freight), the product will struggle to survive ads, shipping and returns. Second, never place your first order above 150 units, whatever discount the supplier waves at you for 500. The discount is cheaper than dead stock.
Before any purchase order, write your per-order math. This is the single spreadsheet that decides whether the business can exist:
If the bottom line is negative on paper, it will be more negative in reality, because reality adds damaged stock, discounts and festival courier surcharges. The full method, with benchmarks by category, is in our guide to D2C unit economics in India.
In my supply chain years at Atomberg, no product proposal reached pricing review without a landed-cost sheet, and I run founders the same way. I ask for the per-order math before I look at the store, the logo or the ads. Nine times out of ten the founder has a beautiful storefront and no answer for what one delivered order earns. That order of attention, math before aesthetics, is the single biggest difference between operators and hobbyists.
Step 4: Marketplace or your own store? Decide with numbers, not opinions
This is the platform decision, and in 2026 the economics have shifted. Amazon India now charges zero referral fee on more than 12.5 crore products priced under ₹1,000 across 1,800+ categories, matching Flipkart's zero-commission move, as reported by Business Standard. Zero referral fee does not mean free: closing fees, weight handling and ad costs still apply.
| Factor | Marketplace (Amazon/Flipkart) | Your own store (Shopify + Razorpay) |
|---|---|---|
| Fees per order | 0% referral under ₹1,000 on covered categories; roughly 2% to 16.5% above, plus closing + weight fees | ~2% gateway + ₹36 to ₹45 shipping + platform subscription |
| Traffic | Built in, you rent it with ads and rank | Zero on day one, you buy or earn every visit |
| Customer data | Stays with the marketplace | Yours: phone, email, repeat purchase |
| Price control | Constant pressure to discount | Full control, bundles, offers |
| Best for | Commodity products, fast cash flow, validation | Brands, repeat purchase, margin over time |
If your product sells under ₹1,000 and competes on price → start on a marketplace, ride the zero-commission window, and let it fund you. If your product is brand-led or bought repeatedly (skincare, coffee, pet supplies) → build your own store, because the second order is where you profit and the marketplace keeps that customer. If you are unsure → list on one marketplace for cash flow and validation while your own store becomes the home for margin and repeat buyers. Never split your energy across three platforms in month one.
The full fee-by-fee comparison, with worked ₹ examples for both routes, is in our Amazon vs Shopify for Indian sellers breakdown.
Step 5: Set up your store (Week 4 to 5 · 3 to 5 days of work)
Now, and only now, build the store. On Shopify, the Basic plan costs ₹1,499 per month on annual billing (about ₹1,994 month to month) per Shopify India pricing, plus 18 percent GST. Add a domain at roughly ₹700 to ₹900 a year. A free theme like Dawn is genuinely enough; paid themes solve problems you do not have yet.
The store needs exactly this much to start converting:
- 3 to 5 products, not 30. Depth of conviction beats breadth of catalogue.
- Photos that answer doubts. Six shots per product: front, back, in-hand for scale, in-use, packaging, and one close-up of texture or build. A phone in daylight beats a lazy studio.
- A product description that sells the outcome in the first two lines and covers size, material and usage below.
- Five legal pages: shipping policy, refund and return policy, privacy policy, terms, contact with a real phone number. Payment gateways check these before approving you, so write them on day one.
Step 6: Set up payments: Razorpay, UPI and the COD question (2 to 4 days)
Shopify Payments does not operate in India, so you connect an Indian gateway. Razorpay is the default choice for most new stores: standard pricing is 2 percent plus GST per successful transaction across UPI, cards, netbanking and wallets, with no setup fee or annual charge, per Razorpay's published pricing. Cashfree and PayU are credible alternatives at similar rates; pick one and move on. Approval needs your GSTIN or proprietorship proof, bank details, and those five policy pages from Step 5.
UPI will quietly dominate your checkout, so make sure it is front and centre. The harder call is cash on delivery. COD still drives a meaningful share of Indian ecommerce orders, especially outside metros, but it carries the RTO risk covered in the next step. The operator position: keep COD on, because switching it off can cost you a third of your orders, but manage it actively with order confirmation and a COD fee rather than pretending the risk does not exist.
Step 7: Shipping: courier aggregators and what delivery really costs
Do not sign a direct contract with a single courier at your volume. A courier aggregator (Shiprocket, Delhivery One, iThink, NimbusPost) gives you 8 to 15 couriers behind one dashboard, discounted rates, label printing, tracking and NDR management. Shiprocket's entry plans start around ₹20 to ₹26 per 500g per Shiprocket's pricing page, but budget with the blended reality: after zone multipliers, COD collection charges (typically ₹40 to ₹50 or about 2 percent, whichever is higher), volumetric weight and GST, a light parcel usually lands at ₹36 to ₹45 forward, and more for far zones.
Three shipping rules that protect margin from day one:
- Weigh and measure your packed parcel before setting prices. Couriers bill on the higher of actual and volumetric weight (length × breadth × height ÷ 5000), and oversized packaging silently doubles freight.
- Price shipping into the product and offer "free shipping" above a threshold that lifts average order value.
- Watch RTO like a hawk. Return-to-origin on COD orders runs 20 to 40 percent for unmanaged stores, and every RTO costs forward freight, reverse freight and repackaging with zero revenue. Our playbook on reducing RTO on COD orders covers the confirmation flows and pincode rules that pull it down.
Step 8: Score your launch, then get your first sales (Week 5 to 6)
Before you spend a rupee on ads, score yourself honestly. According to the Launch Readiness Score™, a store is ready to spend on marketing only when demand proof, margin and operations all clear the bar, because ads amplify whatever exists, including losses.
Launch Readiness Score™: ten checks across three gates. Demand (validated intent from strangers, a price the market accepted, a sharp answer to "why you over Amazon's top listing"), Margin (landed cost sheet done, 3x price cover, positive net per order after CAC and RTO assumptions), Operations (GST and gateway live, aggregator connected, packaging drop-tested, policies published). Score 8 or above: spend on ads. Score 6 to 7: fix the gaps first. Below 6: you are not launching, you are gambling.
- Demand validated with strangers' money or intent, not compliments from friends
- Unit economics sheet shows positive net per order at your realistic CAC
- GST position confirmed with a CA; GSTIN live if marketplaces or inter-state shipping apply
- Current account opened in the business name and linked to the gateway
- Supplier locked with a paid sample test and a first order of 150 units or fewer
- Store live with 3 to 5 products, six photos each, and all five policy pages
- Razorpay live and tested with a real ₹10 UPI transaction end to end
- Courier aggregator connected, one test parcel shipped to a friend in another state
- COD rules set: confirmation flow on, COD fee or prepaid discount configured
- ₹10,000 to ₹25,000 ring-fenced as learning budget for the first ad tests
Then earn the first orders in this order. First, 10 to 20 sales at full price to your extended network, WhatsApp and Instagram DMs, full price because a discount buys politeness instead of truth. Second, organic content in the communities where your buyer already scrolls: Instagram reels, relevant Facebook and WhatsApp groups, local sellers' pages. Third, a small Meta ads test at ₹300 to ₹500 per day for two weeks, judged purely on cost per purchase against the margin sheet from Step 3. Treat the first ₹10,000 as tuition for learning your real CAC; our guide to Meta ads for D2C brands in India covers structure, budgets and kill rules.
A realistic 45-day launch timeline
Put together, the roadmap compresses into six weeks of focused evening-and-weekend work:
- Days 1 to 10: marketplace autopsy, 20 buyer conversations, landing page test. Decision gate: proceed, pivot or kill.
- Days 8 to 15: proprietorship papers, GST application, current account.
- Days 10 to 25: supplier samples, unit economics sheet, first purchase order.
- Days 22 to 32: store build, photos, policies, Razorpay approval, aggregator setup.
- Days 30 to 45: Launch Readiness Score, network sales, organic push, first ad test.
Founders who launch in 45 days did not work harder than the ones still "preparing" at month four. They sequenced the same tasks in an order where each step de-risked the next.
Your next action today
Do not open Shopify today. Open Amazon. Pull up the top 10 listings for the product you want to sell, read the last 50 reviews on each, and write down three complaints that repeat. That is the first pass of the Founder Decision Loop™, it costs nothing, and it will teach you more about your market in 90 minutes than a month of theme customisation. Everything else in this roadmap builds on what you find there.
If you'd like the complete execution system, calculators, SOPs, templates and operating frameworks behind this process, continue inside D2C Acquisition.Lab.
