Shopify vs custom e-commerce: real numbers for SMBs.

Shopify powers over 4.6 million stores and processes roughly 10% of US e-commerce. Most SMBs that leave Shopify for a custom stack regret it within 18 months. That is not an argument against ever leaving — it is an argument for knowing exactly what problem you are solving before you do.

The three real choices in 2026

SMB commerce builds in 2026 collapse to three architectures. There are permutations within each, but for a decision-maker, these are the slots:

1. Standard Shopify (Liquid theme): Shopify hosts everything, you customize a theme, your team manages products and content in the Shopify admin. The default.
2. Headless Shopify: Shopify is the back office; a custom Next.js storefront reads products via the Storefront API. Checkout still uses Shopify (PCI offload + Shop Pay). The storefront is yours.
3. Custom commerce: Next.js + Stripe + your own Postgres. Total control, full stack ownership, no Shopify in the loop. Built when the business logic does not fit Shopify at all.

Revenue threshold: which architecture fits where

What standard Shopify actually gives you

It is easy to underrate Shopify because it is everywhere. The list of things you get for USD 39/month and a weekend of theme work:

• PCI-DSS Level 1 compliance handled. Your servers never touch card data.
• Shop Pay one-click checkout — typically 1.5–2x conversion on repeat buyers across the Shop Pay network of 100M+ shoppers.
• Tax calculation in 200+ jurisdictions via Shopify Tax, plus marketplace tax remittance where applicable.
• Fraud screening on every order. Catches a meaningful share of chargeback risk you would otherwise eat.
• Inventory across multiple locations, with smart routing rules.
• A massive app ecosystem — Klaviyo, Judge.me, Recharge, Loox, the full B2B and DTC tooling stack.
• Themes — paid and free — built by people who know what converts in commerce.

Replicating that with a custom stack is real work. Even at agency rates, you cannot rebuild it for less than USD 60,000 unless you cut corners on something that will bite you.

Where Shopify starts to hurt

Standard Shopify breaks down in five recognizable ways:

1. App sprawl

Most established Shopify stores run 10–18 apps. Each app injects JavaScript, often into the global scope, often on every page. The cumulative weight tanks Lighthouse scores, which tanks Core Web Vitals, which tanks both conversion and SEO. The fix is rarely "a better theme" — it is "cut six apps and write the missing 200 lines in theme code."

2. Content depth

Shopify is great for products. It is mediocre for editorial. If your brand depends on long-form content — guides, lookbooks, magazine sections, technical references — Liquid sections start to feel cramped. Headless or hybrid (Shopify + Sanity for content) opens up real freedom.

3. B2B with negotiated pricing

Shopify B2B (on Plus) handles wholesale price lists and net terms well. It does not handle deal-specific pricing, custom contracts, or approval workflows. For those, you build alongside or instead of Shopify.

4. Subscription complexity beyond Recharge

Recharge covers most subscription cases. It does not cover: usage-based billing, mid-cycle upgrades with proration, multi-product subscription bundles with per-item pause, B2B subscriptions with approval flows. Stripe Billing handles all of those — and Stripe Billing only plays cleanly with custom checkout.

5. Cost at high volume

Shopify takes a percentage of every transaction beyond the platform fee. Above USD 5M ARR, that processing margin starts to dwarf the cost of a Stripe-direct custom stack. The math gets compelling somewhere between USD 5M and USD 15M ARR depending on category.

What headless Shopify changes

Headless Shopify is a middle path more SMBs should consider. Architecturally: Shopify stays as the back office. A Next.js application reads the product catalog through Storefront API, renders pages at the edge, and hands off to Shopify Checkout for the payment step.

You get:

• Page speeds Liquid cannot match. Vercel edge + static product pages + dynamic cart state = sub-1-second first paints worldwide.
• Design freedom. The storefront is React. Animations, layouts, interactions limited only by what you can imagine.
• Content layer flexibility. Sanity, Contentful, Payload, or Notion-as-CMS feeds editorial alongside Shopify products.
• AI-native architecture. JSON-LD, semantic HTML, fast crawl times — your storefront is structured for GPTBot, ClaudeBot, and Google's AI Overviews.

You give up:

• Some Shopify apps that depend on Liquid hooks. Most have headless equivalents; some do not.
• Easy theme switching. Your storefront is custom code now.
• Initial cost. USD 14,000–32,000 vs USD 5,500–14,000 for a Liquid rebuild.

When fully custom is the right call

Fully custom commerce (no Shopify in the stack) is a niche choice. It earns its keep when:

• Your business is more SaaS than store — usage billing, seats, tiered access.
• You operate in a vertical where Shopify's ToS forbids you (some regulated wellness, firearms, certain financial products).
• You sell to enterprises with bespoke contracts and your sales team needs CRM-integrated quoting.
• You are scaled enough (USD 10M+ ARR) that processing and platform fees are a material line item.

Under those conditions, custom is the cheaper long-term call. Above them, you are usually overpaying for engineering.

When custom backfires

• No plugin ecosystem. Every feature a Shopify app delivers in a day (reviews, subscriptions, loyalty, returns) takes weeks to build and test from scratch. Your engineering team spends time on commodity features instead of your actual moat.
• Hard to hire for."Knows our custom Rust checkout logic" is a narrower hire than "knows Shopify Functions." Talent cost and risk both go up.
• Migration cost is high in both directions. Getting off Shopify: USD 15,000–40,000 engineering + data migration risk. Getting back onto Shopify after a failed custom build: same investment, plus rework. Budget 6 months of disruption for a full replatform.
• You lose Shopify's legal coverage. PCI compliance, chargeback management, tax remittance — Shopify handles these by default. Custom means you own the liability.

The five-shape decision table

Common SMB mistakes

• Leaving Shopify too early. The replatform cost eats 18 months of margin. Most stores should rebuild on Shopify first and prove they have hit a real ceiling before going custom.
• Adding apps instead of editing theme code. Every app is a tax on page weight. After app number five, you should have a developer on retainer auditing the JavaScript load.
• Treating headless as a status symbol. Headless is a tradeoff, not a tier. If your team cannot edit the Next.js code base, you have just made your storefront harder to maintain and harder to staff.

How we approach the decision

In an IF Digital discovery sprint, we pull your last 90 days of Shopify Analytics, run a Lighthouse audit of your current storefront, list every app you run, and benchmark your conversion against the category median. The decision falls out of the data in most cases: Liquid rebuild fixes the next 18 months for most established stores; headless is worth the price tag when content depth is a strategic moat; custom is rare and we will say so when it is the answer.

Bottom line

Shopify is not a beginner platform — it is the right platform for the vast majority of SMB stores regardless of stage. Most "we outgrew Shopify" problems are actually "we have too many apps" problems, solvable with a Liquid rebuild at a fraction of the cost of replatforming.

Headless earns its price when you need content-brand depth or conversion-critical page speed at scale. Fully custom earns its price when the business logic genuinely does not fit Shopify's model. Those are specific, identifiable conditions — not a matter of preference or ambition.

Bring us your store URL and a guess at what is broken. We will tell you whether it is the platform, the theme, the apps, or your merchandising — before you spend anything.