If you are a US D2C founder or an ecommerce ops lead, the custom ecommerce vs marketplace decision shapes the next three to five years of your margin structure, brand equity, and engineering bill. This guide walks through the real three-way choice: Amazon-style marketplaces, Shopify-style platforms, and custom headless builds, with 3-year TCO scenarios at $1M, $10M, and $100M GMV.
We are not going to tell you to always build custom. For most brands under $10M GMV, that would be wrong. What we will do is put numbers next to each path, call out the hybrid strategies that most successful DTC brands actually run, and give you a 6-question framework you can take into your next board meeting.
Custom ecommerce vs marketplace: the three-way decision
When founders ask us whether they should build a custom storefront, they usually skip over the real decision. There are three paths, not two:
- Path A — Marketplace-first. You list on Amazon, Walmart Marketplace, TikTok Shop, Etsy, or some combination. You pay take rates in exchange for demand. You do not own the storefront, the checkout, or the customer data.
- Path B — Platform-hosted. You run on Shopify, BigCommerce, or WooCommerce. You own the URL and most of the customer relationship. You rent the runtime and pay an app ecosystem to fill the gaps.
- Path C — Custom / headless. You build on Next.js Commerce, Medusa, Saleor, Commerce Layer, or commercetools, with Stripe for payments and your own engineering team to maintain it. You own everything, including the bugs.
Most brands at scale run A and B in parallel. A few run B and C in parallel. Very few run C alone. The question is which mix, at which GMV, with which team.
Marketplace economics: what Amazon actually costs
Amazon is the default for most US brands selling physical goods because demand is already there. But the unit economics are unforgiving. On a private-label SKU selling at $25:
- Referral fee: typically 15% of the sale price — $3.75 gone before anything ships.
- FBA fulfillment: roughly $3.00-$5.50 per unit for standard-size items, depending on weight and dimensions (higher in 2026 after the latest Amazon fee schedule).
- Storage fees: monthly, plus Q4 surcharges, plus long-term storage penalties if inventory ages.
- Advertising (Sponsored Products): most categories now require 8-15% ACOS just to stay on page one.
- Returns, reimbursements, and removal fees: another 2-4% drag on category-dependent SKUs.
All-in, a typical Amazon FBA private-label seller loses 30-45% of gross revenue to Amazon before cost of goods. FBM (Fulfilled by Merchant) lowers fulfillment cost but raises your ops load and can hurt Buy Box eligibility.
The hidden costs nobody lists
- Account suspension risk. One policy change, one hijacker, one bad batch of reviews, and your listing goes dark. Founders who built 100% on Amazon have lost their business in 48 hours.
- Zero customer data. You get an obfuscated order ID. No email, no phone, no way to retarget or build loyalty without Amazon DSP.
- Brand dilution. Your product ships in an Amazon box. Unboxing, packaging, inserts — none of that builds your brand.
- Price compression. The Buy Box algorithm rewards lower prices. Your competitors are one click away on the same page.
Walmart Marketplace, TikTok Shop, and Etsy have similar structural issues with different numbers. TikTok Shop is the most interesting new channel in 2026 — strong for impulse and discovery, brutal on margin and returns.
Platform economics: what Shopify actually costs at scale
Shopify is the default for US DTC brands because it gets you to live in under a week and scales cleanly into the low eight figures. But it is not free, and the app stack is where the real bleed happens.
Shopify subscription tiers (2026)
- Basic: $39/mo, 2.9% + 30¢ card rate, 2.0% third-party payment fee.
- Shopify: $105/mo, 2.6% + 30¢, 1.0% third-party fee.
- Advanced: $399/mo, 2.4% + 30¢, 0.5% third-party fee.
- Shopify Plus: starting around $2,300/mo, plus transaction fees and enterprise add-ons. Required in practice above roughly $1-2M GMV for checkout customization, multi-store setups, and B2B features.
The app ecosystem tax
A real DTC Shopify stack in 2026 looks like this:
- Klaviyo for email and SMS: $200-$2,500+/mo depending on list size.
- Gorgias or Zendesk for support: $100-$1,000+/mo.
- Yotpo or Okendo for reviews and loyalty: $200-$1,500/mo.
- Loox or Judge.me for photo reviews: $50-$500/mo.
- Recharge or Skio for subscriptions: 1-2% of subscription GMV.
- Malomo, Route, or Shop Pay for post-purchase: $100-$500/mo.
- Rebuy, Nosto, or Tapcart for personalization / mobile: $500-$3,000/mo.
At $10M GMV, most DTC brands spend $2,000-$5,000/month on Shopify apps alone, on top of the platform fee. Shopify Markets adds international capabilities (currency, language, duty) but still charges on international GMV.
Custom headless economics: what it really takes
Custom or headless does not mean "rebuild Shopify from scratch." It means you use Next.js Commerce, Medusa, Saleor, commercetools, or Commerce Layer for the commerce engine, Stripe for payments, a CMS like Sanity or Contentful for content, and your own team to stitch it together.
- Initial build: $100k-$1M+ depending on scope. A minimal headless storefront with a clean checkout, PDP, cart, and one CMS integration is typically $120k-$250k. A full replatforming with loyalty, subscriptions, multi-region, B2B portal, and custom merchandising runs $400k-$1M+.
- Ongoing engineering: $30k-$150k/mo for a 3-8 person team (frontend, backend, DevOps, QA). This is the number most founders underestimate.
- Infrastructure: Vercel Pro/Enterprise, AWS, observability (Datadog, Sentry), image CDN (Cloudinary, imgix). Typically $2k-$20k/mo.
- PCI scope: outsourced to Stripe or Adyen — you never touch a PAN, which removes most compliance burden.
The payoff: no platform take rate beyond payment processing (2.2-2.9% on Stripe), total control over checkout conversion, merchandising logic, site speed, and content/commerce integration. Brands crossing $25-50M GMV with strong engineering culture usually find this math works.
3-year TCO scenarios
Below are rough 3-year total cost of ownership estimates across three GMV tiers. Numbers assume US-based DTC with 50-55% gross margin pre-platform fees and normal DTC marketing spend (not included). All USD.
Scenario 1: $1M GMV per year
| Option | Platform cost (3yr) | Take rate / processing | Apps / integrations | Engineering | 3-yr TCO |
|---|---|---|---|---|---|
| Amazon FBA only | $0 | ~$1.0M (33% all-in) | $15k (tools) | $0 (agency $30k) | ~$1.05M |
| Shopify Basic/Shopify | $3.8k | ~$90k (3% blended) | $25k | $30k (freelance) | ~$150k |
| Shopify Plus | $85k | ~$75k | $60k | $50k | ~$270k |
| Custom headless | n/a | ~$75k | n/a | $250k build + $720k ops | ~$1.05M |
At $1M GMV, Shopify wins on almost every axis. Custom is hard to justify unless you are raising growth capital specifically to build a tech moat.
Scenario 2: $10M GMV per year
| Option | Platform cost (3yr) | Take rate / processing | Apps / integrations | Engineering | 3-yr TCO |
|---|---|---|---|---|---|
| Amazon heavy (80%) | $0 | ~$9.9M (33% on $24M) | $75k | $150k | ~$10.1M |
| Shopify Advanced | $14k | ~$810k | $180k | $300k | ~$1.3M |
| Shopify Plus | $85k | ~$750k | $360k | $600k | ~$1.8M |
| Custom headless | n/a | ~$750k | $60k | $400k build + $2.7M ops | ~$3.9M |
At $10M GMV, Shopify Plus is the sweet spot for most brands. Custom makes sense only if you have a merchandising or content-commerce thesis the platform genuinely cannot deliver.
Scenario 3: $100M GMV per year
| Option | Platform cost (3yr) | Take rate / processing | Apps / integrations | Engineering | 3-yr TCO |
|---|---|---|---|---|---|
| Shopify Plus | $240k | ~$7.5M | $1.5M | $1.8M | ~$11M |
| Shopify Plus + extensive custom | $240k | ~$7.5M | $1.5M | $5.0M | ~$14.2M |
| Custom headless | n/a | ~$7.0M | $200k | $800k build + $5.4M ops | ~$13.4M |
At $100M+ GMV, the math gets close. Custom wins on conversion lift potential (a 1% CVR gain = $1M/yr), on content-commerce flexibility, and on international edge cases. Shopify Plus wins on speed to ship and on hiring: Shopify engineers are easier to find than headless commerce generalists.
Real examples, briefly
- Allbirds ran on Shopify Plus for years before going public and still uses it for the core storefront.
- Warby Parker built a largely custom stack early — the eyewear try-on and prescription flows did not fit any off-the-shelf platform.
- Everlane scaled on Shopify Plus and invests heavily in content-commerce integration on top.
- Most CPG brands (Liquid Death, OLIPOP, Magic Spoon) run Shopify Plus for DTC and lean heavily on Amazon and retail for volume.
- Luxury and fashion (off-white, select direct-to-consumer Kering brands) tend custom or Salesforce Commerce Cloud / commercetools for full brand control.
The pattern: category fit, margin structure, and merchandising complexity drive the decision more than GMV alone.
Hybrid strategies that actually work
Almost no successful US DTC brand is pure single-channel. The common patterns:
- Own site primary, Amazon as overflow: push new launches, premium SKUs, and subscription exclusively on your site. Use Amazon for established SKUs where customers already search.
- High-margin on own, low-margin on marketplace: full-price seasonal on Shopify, clearance and liquidation on Amazon.
- B2C on Shopify, B2B direct: Shopify Plus B2B or a custom B2B portal (NetSuite SuiteCommerce, commercetools B2B) for wholesale buyers with credit terms and negotiated pricing.
- DTC on Shopify + headless storefront for content: commerce stays on Shopify (via Storefront API), the marketing site goes headless on Next.js for SEO, speed, and editorial flexibility. Very common at $10-50M GMV.
Re-platforming risk is real
Changing platforms costs more than founders expect. Typical ranges:
- Shopify to Shopify Plus: 4-8 weeks, $30-100k services.
- Shopify to headless: 3-9 months, $150-500k build, plus ongoing engineering cost.
- Legacy Magento/BigCommerce to Shopify Plus: 4-12 months, $200k-$1M.
- Full custom rebuild: 6-18 months, $500k-$3M+ including lost productivity.
The cost you do not see in the invoice: SEO equity loss if redirects are not perfect, customer data loss if export/import is sloppy, and 2-6 months of conversion dip as the new site is tuned. Budget for a dedicated technical SEO lead during any replatforming. See our AI for ecommerce features breakdown for what to prioritize if you are modernizing during the move.
A 6-question decision framework
Answer these six questions honestly before picking a path.
- What is your current and projected GMV over 3 years? Under $500k, stay on marketplaces or Shopify Basic. $500k-$10M, Shopify or Shopify Plus. $10M-$50M, Shopify Plus is usually right. $50M+, headless is on the table.
- How margin-sensitive is your category? Low margin (30-40% gross): marketplace fees hurt more, platform-hosted is usually better than marketplace-heavy. High margin (60%+): you can absorb more platform cost and should prioritize brand control.
- How central is brand experience to your thesis? If your product is a commodity, marketplaces are fine. If your product is a brand (beauty, fashion, lifestyle, supplements), own the customer relationship.
- What is your team's engineering depth? No in-house engineers: Shopify, period. 1-3 engineers: Shopify with custom theme work. 5+ engineers with commerce experience: headless is viable.
- How international will you be in 3 years? Heavy international: Shopify Markets or headless with localized storefronts. US-only: simpler stack wins.
- Is your merchandising logic standard or unusual? Standard catalog: Shopify handles it. Configurable products, B2B pricing tiers, subscription with pause/skip, try-before-you-buy, custom quoting: headless or heavy Shopify customization.
No single answer wins. Consistency across the six is what tells you which path. If five of six point to Shopify Plus and one points to headless, pick Shopify Plus and solve the outlier with a headless microsite.
Where nearshore engineering changes the math
Custom builds get cheaper, not easier, with nearshore teams. A US-based senior full-stack engineer runs $180-$280k/yr fully loaded in 2026. A senior Brazilian engineer through a nearshore partner runs roughly $85-$140k/yr, with timezone overlap across all US zones (Brazil sits 1-3 hours ahead of US time) and strong Next.js, Node, and headless commerce experience clustered in Sao Paulo, Florianopolis, and Curitiba.
At FWC, our typical nearshore ecommerce engagement is a 3-5 person pod (frontend, backend, DevOps) working against a US product manager, with daily standups in US morning hours. A headless build that quotes $600k in the US lands closer to $280-380k nearshore with comparable senior-engineer output. If you want the full breakdown, read our nearshore web and ecommerce development guide and the custom software guide for US enterprises.
When to stay on Shopify, when to go custom
Stay on Shopify / Shopify Plus if: you are under $25M GMV, your category fits standard commerce patterns, your team does not have 5+ engineers, or you are optimizing for speed of iteration over long-term unit economics.
Go custom / headless if: you are over $25-50M GMV, your merchandising requires logic Shopify cannot express cleanly, your brand experience demands content-commerce integration the platform cannot deliver, you have strong engineering leadership, or you are building a B2B or marketplace-style product on top of commerce primitives. Cost estimates for a custom ecommerce or mobile commerce app are in our 2026 US app development cost guide.
Ready to decide?
If you are weighing custom ecommerce vs marketplace and want a second opinion grounded in real TCO numbers, we can help. At FWC we have built DTC storefronts, B2B portals, and custom commerce integrations for US and LATAM brands, and we are comfortable telling you when Shopify is the right answer and we are the wrong team for the job. Start with the 10 questions to ask before hiring a software company and then get in touch or request a quote with your GMV, current stack, and 12-month roadmap.
The custom ecommerce vs marketplace choice is not a binary. It is a portfolio decision that should match your GMV, margin, brand thesis, and team. Get the framework right, let the TCO math tell you which platform, and revisit every 18-24 months as your business evolves.