Shopify vs custom Stripe for UK e-commerce: when to pick which
Cost-per-order, integration headaches, and the £39/mo break-even point that flips Shopify into a worse deal.
For a UK SMB doing under 1,000 orders / month, Shopify's Basic plan at £19 / month (annual billing) is the right answer. Above 1,000 orders / month, the maths flips — and most owners don't notice because the fees are buried in the transaction line.
Shopify true cost
Plan: £19 / month annual or £25 / month rolling. Transaction fee on Shopify Payments: 2.0% + 25p as of the early-2026 schedule (it has crept up from the 1.7% rate quoted in older posts — Shopify revised UK domestic card processing in late 2025). Apps: the average client we audit pays £80–£140 / month for apps that should ship for free (reviews, abandoned cart, upsells).
A custom Stripe checkout
Hosting: £15 / month. Stripe: 1.4% + 20p for UK consumer cards. Apps: built-in. No transaction fee on top, no per-product limits.
The break-even
At 950 orders / month with an average basket of £55, the Shopify stack costs £1,162 / month at the older 1.7% rate, more like £1,320 / month on the 2026 schedule. The Stripe stack costs £765. The £400-£550 / month delta funds a redesign every quarter.
When Shopify is still the right call
Multi-warehouse stock sync, Shopify POS in physical stores, or international tax compliance you don't want to build yourself. Then £19 / month + transaction fees is a bargain.
When Stripe wins
High-AOV, low-volume product (one-off services, deposits, custom quotes). Subscription with annual prepay. Anything where the per-order fee is the dominant cost.
The full maths, written out
Take a UK B2C ecommerce business doing 1,200 orders / month at £55 average order value, 100% Shopify Payments, on the Basic plan. Monthly revenue: £66,000. Shopify plan: £19. Transaction fees at 2.0% + 25p: £1,320 in percentage + £300 in flat = £1,620. Apps (reviews, upsell, loyalty, SMS): £110. Theme licence amortised: £8. Total Shopify cost: £1,757 / month, or 2.66% of revenue. Same volume on a Next.js + Stripe build: hosting £20, Stripe at 1.4% + 20p = £924 + £240 = £1,164, plus £25 / month for transactional email and a logging stack. Total: £1,209 / month, or 1.83% of revenue. The 83 basis-point gap is £548 / month, £6,576 / year. At year three you have funded a redesign, a paid-search budget, or simply taken the cash home.
What you give up by leaving Shopify
Be honest about this: leaving Shopify costs you a working admin UI, a 10,000-app ecosystem, official Shopify POS, an out-of-the-box Shop Pay accelerated checkout, the fraud-protection engine, the discount engine, and the abandoned-cart automation. A custom Stripe build replicates the checkout in about 80 hours, replicates the discount engine in another 40, and asks the founder to either build or buy each of the other capabilities. For a five-product brand with a clean ordering flow, the gap is invisible. For a 400-SKU fashion brand running flash sales and bundles, the gap is large enough that staying on Shopify even at 2.66% is the rational choice.
The hybrid that quietly wins for most owners
Run Shopify for the storefront and inventory; run a custom Stripe checkout for high-ticket bespoke orders that don't fit the SKU model (custom builds, services, deposits, B2B). Most of our six-figure clients end up here — they use Shopify's strengths where Shopify is irreplaceable and route the high-margin custom work through a dedicated Stripe flow that does not pay the per-order tax. Implementation is a single shared customer record across both systems (Klaviyo or HubSpot makes a fine spine) and a unified analytics view in GA4.
App audit: what most stores are paying for that they shouldn't
Five apps we strip from Shopify audits, in order of waste: (1) Loox / Judge.me Premium — Shopify's built-in Reviews app now does 90% of what they do, free; (2) Klaviyo on a small list — under 1,000 contacts the free Shopify Email is genuinely enough; (3) Bold Upsell — Shopify's native post-purchase upsell shipped in 2024 and matches it; (4) Privy popups — most themes ship an email capture; the £25 / month is buying lipstick; (5) most "loyalty" apps for stores doing under £20k / month — the maths on points programmes does not work below that volume.
Stripe fees that catch people out
Stripe's 1.4% + 20p is for UK consumer cards (domestic, present). The card fee schedule has more bands than the website implies. Non-EEA cards: 2.9% + 20p. European Economic Area commercial cards: 2.5% + 20p. American Express: 2.5% + 20p. Currency conversion: 2% on top if you settle in GBP from a non-GBP charge. Disputes: £15 each (refundable if you win). Instant payouts: 1%. Stripe Tax (if you turn it on): 0.5% per transaction. For a UK-only consumer store, the headline 1.4% is accurate; for a cross-border B2B store the blended rate creeps closer to 2.4%.
When the comparison should not even happen
If you sell on Amazon, eBay or TikTok Shop as well, the Shopify admin's inventory sync is worth the platform tax on its own. If you are pre-product-market-fit and changing the offer weekly, do not waste 80 hours on a custom Stripe build — Shopify lets you change everything in an afternoon. The break-even argument only matters once the offer is stable, the orders are repeatable, and the per-order economics will compound for 24+ months.
Migration: moving from Shopify to a custom Stripe stack
For owners who decide the maths has flipped, here is the actual sequence we run on a migration. Week one: full export from Shopify (products, customers, orders, blog posts, redirects), Search Console crawl of every URL the store currently ranks for, full inventory of apps in use and their data-out paths. Week two: the new Next.js + Stripe build on a staging URL, with the catalogue imported, the checkout flow live, the customer accounts migrated (passwords reset on first login via magic link rather than hash-import — Shopify does not export the password hashes and that is a feature, not a bug). Week three: 1:1 redirect map from every old /products/[slug] and /collections/[slug] to the equivalent new URL, plus catch-all rules for the long tail; full QA pass on the checkout edge cases (saved cards, abandoned cart recovery, refund flow, partial fulfilment); content review of any blog posts that ranked well. Week four: launch on a Tuesday morning, DNS swap with low TTL, Search Console fresh sitemap submitted, marketing email cadence resumed. Day 7, 14, 30 health checks confirm rankings and conversion held.
What we have seen go wrong on Stripe builds
Two failure modes worth pre-empting. First: forgetting Strong Customer Authentication (SCA) on the initial card-on-file capture for subscriptions. Stripe handles SCA correctly when you use the Setup Intent API on the initial save, but if the developer cut a corner and used a raw PaymentMethod attach, the first recurring charge will silently fail on UK and EEA cards under PSD2. We test the SCA path explicitly. Second: not configuring Stripe Tax for the right jurisdictions, or configuring it and forgetting that the per-transaction Tax fee (0.5%) is not zero. For UK-only businesses below the VAT threshold, Stripe Tax is unnecessary and the fee is pure waste; for businesses crossing the threshold or selling into the EU, Stripe Tax is the cheapest option versus Avalara or TaxJar but the maths needs to be in the model.
Recurring revenue: where Stripe Billing earns its keep
Stripe Billing is a separate product layered on top of Stripe Payments and it is where subscription-led businesses get the biggest gap versus Shopify Subscriptions. Native Stripe Billing handles: complex pricing models (graduated, volume, package, tiered), prorations on plan changes mid-cycle, custom dunning sequences with configurable retry windows, automated tax handling, invoice generation, and a hosted customer portal where the subscriber can manage their own plan. The combination removes 60-80% of the integration work that subscription businesses pay developers for. Shopify Subscriptions is fine for simple SKU-and-frequency offers (every 30 days, every 60 days) but lacks the pricing-model depth and the API surface area. For any subscription business above 100 active subscribers we default to Stripe Billing on a custom front-end.
Sales tax and VAT in 2026
A note on the regulatory layer because it has shifted. UK VAT registration threshold is £90,000 of taxable turnover (raised from £85,000 in April 2024); below that you charge no VAT but cannot reclaim input VAT. EU OSS (One-Stop Shop) registration is required for any UK business selling more than €10,000 / year in physical goods to EU consumers — Stripe Tax handles the rate calculation; you still need to register with HMRC for OSS and file the quarterly return. US sales tax has its own complexity (nexus thresholds per state, marketplace facilitator laws); if you are selling into the US, either Stripe Tax (cheapest) or TaxJar (better UI) handles the per-state rules. None of this is uniquely a Stripe-versus-Shopify decision; both stacks need the same tax layer once you cross a threshold.
A decision matrix for the hesitant owner
For owners trapped between the two options, here is the heuristic we walk clients through. If your monthly order volume is below 400, stay on Shopify Basic — the platform fee is small in absolute terms and the time savings on day-to-day operations are real. If your monthly order volume is between 400 and 1,200 and your offer is stable, run the maths on both stacks and pick the lower number; both are reasonable. If your monthly order volume is above 1,200 with a stable offer that will not change materially for 18 months, move to custom Stripe — the per-order fee gap compounds into real money. If you are doing more than £500k / year through Shopify Payments, also consider Shopify Plus negotiated rates, which can close the gap to 0.6%-0.8% on the percentage component and change the picture. The conversation we never want to have is the one where an owner has been paying 2.5% of revenue for three years without knowing the alternative existed.
Final word on the choice
Neither platform is the right answer in every case. Both are mature, both are well-engineered, both serve the businesses they fit. The cost of getting the choice wrong is not catastrophic — migrations are routine — but the cost of never running the maths in the first place is real, compounding, and entirely avoidable. If nothing else in this post sticks, the takeaway is: pull your last twelve months of orders, multiply the volume by the headline rates, and look at the number.