An ecommerce build in 2026 runs from a few hundred dollars for a theme setup to $80,000 and up for headless, and takes two weeks to nine months. The platform choice matters less than the two numbers underneath it: your monthly order volume, and how unusual your pricing is. Most brands should launch on Shopify or WooCommerce and go custom only when checkout logic, wholesale pricing or ERP depth starts fighting the platform. What follows is the cost math, the compliance deadlines already in force, an ownership clause, and a test matrix for your vendor.
Scope decisions are the cheapest decisions you will make
If your store needs subscriptions, custom quoting, location-based inventory, bundled offers, wholesale pricing, or integration with a CRM or ERP, those decisions should shape the build from day one.
Founders rush scope because design feels like progress. Answer these before accepting a quote — each moves the price, and a vendor who does not ask is guessing:
- How many SKUs at launch, and in eighteen months? Under 500 and over 5,000 are different builds.
- How many price lists? Wholesale tiers or quoting is a different data model from one retail price.
- One-time, subscription, or both? In India that carries regulatory consequences and is expensive to defer.
What a build costs in 2026, and what each band buys
No reliable public dataset exists for Indian agency pricing; the blogs publishing ranges quote their own rate cards. Treat the rupee column as a sanity band, not a quotation.
| Band | What you get | USD | Indicative INR | Timeline |
|---|---|---|---|---|
| Theme setup | Paid theme, your branding, catalog loaded, standard checkout | $0–$5,000 | ₹40,000–₹1.5 lakh | 2–4 weeks |
| Customized theme | Theme reworked to your merchandising, custom templates, 3–6 integrations | $5,000–$25,000 | ₹1.5–₹6 lakh | 4–10 weeks |
| Custom storefront | Your business logic: ERP/CRM integration, B2B pricing, admin tooling | $25,000–$80,000 | ₹6–₹25 lakh | 2–4 months |
| Headless / custom | Decoupled frontend or an engine written for you; multi-market, quoting | $80,000–$250,000+ | ₹25 lakh+ | 4–9 months |
USD bands: BigCommerce, "Ecommerce cost" (vendor-published, read 8 August 2026). INR figures are indicative equivalents, not a price list.
Development is rarely what makes a project late — product data, photography, copy and unstaffed testing on your side are. A vendor who resists a dated client-deliverables schedule has not planned the project.
Shopify, WooCommerce or custom: run the take-rate math first
Shopify Payments is not available in India, so an Indian store uses a third-party gateway and pays Shopify's surcharge on top of the gateway's rate — two take rates, stacked. At ₹10,00,000 monthly revenue on Basic, with Razorpay at 2% + GST:
- Subscription ₹1,994 + Shopify surcharge at 2% (₹20,000) + Razorpay 2% plus 18% GST, i.e. 2.36% (₹23,600)
- ₹45,594 a month — a blended 4.56% of revenue, or ₹5.47 lakh a year.
If you are GST-registered the 18% is normally recoverable as input tax credit, putting the net nearer 4.2%. The ₹20,000 line is platform rent: it scales with every rupee and buys nothing. Basic to Grow pays for itself at about ₹5.45 lakh monthly GMV; Grow to Advanced not until roughly ₹56.8 lakh, so at ₹10 lakh Advanced costs more than Basic.
| Shopify | WooCommerce | Custom / headless | |
|---|---|---|---|
| Best fit | Up to ~₹25–30 lakh/mo, under ~2,000 SKUs, one price list | ~₹5 lakh–₹1 crore/mo, custom pricing or tax logic, dev on retainer | Above ~₹50 lakh/mo, or B2B tiers, quoting, ERP depth |
| Build & timeline | ₹40,000–₹6 lakh, 2–10 weeks | ₹1.5–₹15 lakh, 1–3 months | ₹25 lakh+, 4–9 months |
| Fixed monthly | ₹1,994 Basic / ₹7,447 Grow / ₹30,164 Advanced / ₹1,75,000+ Plus, plus apps | Hosting ~₹3,000–₹15,000, plus plugin licenses | Infrastructure ~₹10,000–₹50,000, plus engineers |
| Take rate | Gateway ~2% + GST plus surcharge 2% / 1% / 0.6% / 0.2% by plan | Gateway ~2% + GST. No surcharge. | Gateway ~2% + GST. No surcharge. |
| All-in on ₹10L/mo | ₹45,594 (4.56%) Basic; ₹41,047 (4.10%) Grow | ~₹28,600 (2.86%); ~₹43,600 with a ₹15,000 retainer | ~₹33,600 (3.36%) before salaries |
| The wall | Checkout logic you cannot change; checkout.liquid and additional scripts on Thank-you/Order-status pages were sunset on 28 Aug 2025, and script tags sunset for non-Plus stores on 26 Aug 2026 | You own patching, security, performance, plugin conflicts | You own everything, including hiring people who understand it |
| Exit | You own the theme code and data, not the platform | Yours; move hosts in a weekend | Yours, and cannot be left unattended |
Shopify India plan prices and the per-plan third-party payment provider fees read from shopify.com/in/pricing on 8 August 2026, monthly billing (annual: ₹1,499 / ₹5,599 / ₹22,680); confirm your own rate in your admin before signing. Razorpay standard pricing 2% plus 18% GST, from razorpay.com/pricing. Deprecation dates from Shopify's checkout.liquid developer documentation.
The honest reading: ₹2.4 lakh a year in surcharge is roughly a whole WooCommerce build, but add a retainer and WooCommerce lands close to Shopify's total. Fees alone rarely justify leaving; checkout logic you cannot express and pricing rules the platform will not model do.
A custom build is not about making things complicated. It is about avoiding limitations that cost you more later. If your team ends up relying on workarounds for inventory syncing, order routing, customer segmentation, or promotions, that technical debt shows up as manual work, errors, and slower growth.
This is why experienced teams look beyond launch. They ask what your store will need in six months, not just what it needs this week. MyMind Studio takes that approach because a rushed build with hidden constraints is not a win. A store should be easy to operate after launch, not just nice to demo before it.
The running cost is the number that compounds
Over three years the build is the smaller number. Budget every line:
- Platform subscription, which steps up as you grow
- Gateway MDR: ~2% + GST in India; 1.5%–3.5% plus a per-transaction fee internationally, plus the platform surcharge where it applies
- Apps and plugins — each a recurring charge and a script on your pages
- Hosting and infrastructure; content and photography, always underbudgeted
- Maintenance, commonly budgeted at 10–20% of build cost a year (a rule of thumb, not a measured figure)
Checkout is an engineering deliverable with a number attached
Baymard Institute puts average documented cart abandonment at 70.22%, from 50 studies. The breakdown matters more than the headline: each cause maps to a build decision.
| Why they abandon | Share | The build decision that fixes it |
|---|---|---|
| Extra costs too high (shipping, tax, fees) | 40% | Landed cost — shipping and tax — at cart entry and on the product page, not at step three |
| Delivery too slow | 20% | A real delivery date on the product page from the courier API, not a static "3–5 days" |
| Did not trust the site with card details | 19% | Tokenized gateway, no card fields in your DOM, visible payment marks, clean checkout scripts |
| Site required account creation | 18% | Guest checkout by default; account claim after the order, using the email already collected |
| Checkout too long or complicated | 17% | A field-count budget in the scope, address autofill, pincode-driven city and state |
| Website errors or crashes | 17% | The test matrix below, plus error monitoring alerting before launch day |
Baymard Institute, updated 22 September 2025.
Put Core Web Vitals in the contract as acceptance criteria
"Fast" is not a specification; Google's thresholds are. LCP at or under 2.5 seconds, INP at or under 200 milliseconds, CLS at or under 0.1, all at the 75th percentile and measured separately on mobile. Write them into acceptance criteria for the product page, collection page and cart, and make final payment contingent on them. Deloitte's study with Google, "Milliseconds Make Millions," found a 0.1-second mobile speed gain lifted retail conversions 8.4% and order value 9.2%, across retail, travel, luxury and lead-generation brands in Europe and the US — 2020 data, so directional.
Published mobile conversion benchmarks disagree, and none of them is your store. Littledata's benchmark of 2,800 Shopify sites puts mobile at 1.2% against 1.9% on desktop; other vendors' panels report a gap of a different size, drawn from a different mix of merchants. Measure your own gap in GA4 rather than budgeting against someone else's sample.
Compliance that is already in force and changes the build
PCI DSS v4.0.1 constrains what runs on a checkout page
Requirements 6.4.3 and 11.6.1 became mandatory on 31 March 2025: an inventory of every script on a payment page with a documented justification for each, integrity verification against unauthorized modification, and a tamper-detection mechanism checking payment-page content and HTTP headers at least once every seven days — or on a longer cycle only if you can defend it with a documented targeted risk analysis. "Just add the pixel to checkout" is no longer a five-minute request.
The European Accessibility Act applies if you sell to EU consumers
Directive (EU) 2019/882 became enforceable on 28 June 2025 and ecommerce is explicitly in scope; the conformance route is EN 301 549, built on WCAG 2.1 Level AA. Retrofitting a finished storefront costs several times what building to it costs, because the work lands in the component layer.
India: saved cards and subscriptions are regulated, so design them in
Since 1 October 2022, RBI's tokenization rules bar merchants and aggregators from storing the card number, CVV or expiry — only the last four digits and issuer name. RBI's e-mandate framework requires authentication at mandate registration, modification and revocation; later debits are AFA-exempt up to ₹15,000 per transaction, raised from ₹5,000 by RBI's circular of 16 June 2022, with higher ceilings since allowed for a few named categories such as insurance premiums, mutual fund subscriptions and credit card bills — check the limit that applies to you before you build the flow. A notification must reach the customer 24 hours before each debit, with an opt-out — a scheduled job, a template, a delivery log, a failure path. "We will add subscriptions later" is expensive here.
Your tax classes may already be wrong
The GST framework revised with effect from 22 September 2025 abolished the 12% and 28% slabs, leaving 5% and 18% plus 40% for luxury and sin goods; a catalog on hardcoded 12% or 28% classes is issuing wrong invoices today. Marketplace sellers: operators collect TCS on taxable supplies, and registration for selling through an ecommerce operator does not follow the ordinary turnover threshold — many sellers must register from their first order, with a narrower exemption for small intra-state suppliers of goods. Confirm your own position with your accountant before launch, because it changes your invoicing.
Integrations and returns belong in the data model, not the roadmap
The integration questions that matter are the failure cases: payment failing mid-order, stock changing between cart and confirmation, a webhook arriving twice. If those answers are vague, you are not ready to build.
A clean integration setup reduces manual work and protects accuracy. It also gives leadership real visibility into sales, fulfillment and margin.
Returns are the integration everyone treats as a policy page. The National Retail Federation's returns research put the online return rate at 17.6% — $247 billion of merchandise bought online and sent back. Roughly one order in six reversing is a core flow: it needs an RMA record with a state machine, partial refunds, restocking, refund routing with a fallback when the original method is dead, and a status support can read. Added later, that means changing the order schema on a live store. (NRF, 2023 consumer returns research.)
Machine readability is now a distribution channel
Product and Offer structured data, AggregateRating, variants, and shipping and return policy nested under Organization make your catalog legible to systems that are not a human with a browser. Google states that providing both on-page markup and a Merchant Center feed maximizes eligibility.
The 2026 reason to care is agentic shopping. The Agentic Commerce Protocol, published by Stripe and OpenAI in September 2025 under an Apache 2.0 license, is implemented in ChatGPT's checkout using a shared payment token, so the agent never sees raw card credentials. Google's competing stack pairs the Universal Commerce Protocol with AP2, announced 16 September 2025 and now being standardized inside the FIDO Alliance, which uses signed mandates. Nobody can tell you today what share of your revenue this will carry, and any vendor quoting you a percentage is guessing. The protocol-agnostic move is cheap: accurate GTINs, real stock status, current prices, shipping and return policy in both feed and markup, and pages rendered in HTML rather than in the browser.
The ownership clause to agree before you pay a deposit
Every agency page says "you own the code." Alone that is worth little. Get these twelve in writing:
- Repository under your organization account, agency as collaborator — not the reverse
- Hosting account in your company's name, on your billing card
- Domain registrar and DNS in your name, with access you have used
- Payment gateway account in your legal entity, settling to your bank
- GA4 property owned by your Google account
- Search Console verified as your property
- Google Merchant Center under your account
- Ad pixels and conversion APIs in your Business Manager
- Theme, plugin and app licenses in your name, transferable
- Full export of products, customers and orders on request
- API keys and environment variables handed over in a password manager
- No clause making continued access conditional on a retainer
A vendor who resists any of these is telling you something.
The pre-launch test matrix to hand your vendor
"We test thoroughly" is not a plan. Ask for this back signed off, with a tester's name and date per line:
- Tax by state and destination country, including zero-rated and exempt products
- Coupon stacking, expiry, minimum-cart rules, one coupon plus one automatic discount
- Partial refund, full refund, refund to an expired card; inventory oversell when two customers buy the last unit in the same second
- Variant edge cases — out-of-stock variant, single-variant product, variant-specific weight
- Guest checkout end to end, post-purchase account claim, password reset
- International address formats — no state field, alphanumeric postcodes — plus COD by pincode and cancellation
- Failed-payment retry; webhook replay — does a repeated order-paid event create two orders
- Order email deliverability with SPF, DKIM and DMARC passing; shipping rates at weight and value boundaries, and at the free-shipping threshold
- Checkout on a throttled 4G connection on a real mid-range Android device
- Purchase analytics firing once with correct revenue and tax; redirects from old URLs; a keyboard-only pass to order confirmation
Frequently Asked Questions
How much does an ecommerce website cost to build in 2026, and what does each band get me?
Roughly $0–$5,000 for a theme setup, $5,000–$25,000 for a customized theme, $25,000–$80,000 for a custom storefront and $80,000–$250,000+ for headless, per BigCommerce's bands. What sets the band is how much of your business logic the platform cannot express — forty products with wholesale tiers costs more than 400 at one price.
How long does it take, and what usually causes the delay?
Two to four weeks for a theme build, four to ten weeks for a customized theme, two to four months for a custom storefront, four to nine for headless. The delay is almost never development — it is product data, photography, copy and unstaffed testing on your side.
Shopify, WooCommerce or custom — is there a revenue level where the answer changes?
Shopify below roughly ₹25–30 lakh monthly GMV with a standard catalog, WooCommerce when you need pricing or tax logic the platform will not model, custom or headless above ₹50 lakh or when B2B tiers and ERP depth are core. If fees are your only reason to leave Shopify, the math says stay.
Why does my Indian Shopify store pay 2% on top of my gateway's charges, and can I avoid it?
Because Shopify Payments is unavailable in India, so you use a third-party gateway and Shopify applies its surcharge — 2% Basic, 1% Grow, 0.6% Advanced, 0.2% Plus. Stacked on Razorpay's 2% + GST that is a blended 4.56% at ₹10 lakh monthly revenue, or ₹5.47 lakh a year. Moving up a plan cuts the surcharge; only leaving Shopify removes it.
Do I own the code, hosting, domain and analytics accounts when the project ends?
Only if you wrote it into the contract — the industry default is that the agency holds several of them. Work through the twelve-item list above and verify each by logging in yourself.
Ready to build an ecommerce system that works as hard as you do? Visit mymindstudio.ai/free-business-growth-audit for a free Business Growth Audit — or talk to the MyMind Studio team about your store.
