Should You Add Subscriptions? What They Actually Cost on Each Platform

Subscriptions get sold to merchants as a retention strategy. They're actually a billing infrastructure problem wearing a retention strategy's clothes, and the difference shows up about four months in, when the failed payments start and nobody owns the dunning process.

Here's a number that frames the honest version. Shopify's own first-party subscriptions app is free, and it sits at 3.7 stars across 711 reviews, with 20% of those being one star^[1]. That's the platform's own product, on the platform it was built for. Subscriptions are hard, and the difficulty is not where the marketing says it is.

This piece works out what recurring billing actually costs on Shopify, WooCommerce, Adobe Commerce, BigCommerce, Swell and Medusa, using published pricing only. Then the regulatory bit that most subscription content ignores entirely, which is that the UK legislated a new subscription regime in 2024 and it changes what your cancellation flow has to look like.

The real cost, compared

Published pricing, accessed 22 July 2026. This is the table I wanted and couldn't find, so we built it from the vendors' own pages.

Route	Platform cost	Subscription cost	On top of gateway?
Shopify + Shopify Subscriptions	From £25/mo[2]	Free[1]	No extra
Shopify + Recharge Starter	From £25/mo[2]	$99/mo + 1.49% + 19c per transaction[3]	Yes
Shopify + Recharge Plus	From £25/mo	$499/mo + 1.34% + 19c[3]	Yes
WooCommerce + Subscriptions	Hosting only	£213/year[4]	No
Swell	$29/mo Starter[5]	Included on every plan[5]	No
Medusa	Self-hosted	No module exists. You build it[6]	Developer time
BigCommerce	From $29/mo[7]	Third-party app	Varies
Adobe Commerce	Licence	Extension or build	Varies

Look at the Recharge row properly, because the "+ 1.49%" is doing a lot of quiet work. That percentage is charged on the subscription transaction in addition to your payment processor's fee^[3]. Recharge says so plainly on their own pricing page: it's "in addition to any other services or solutions your store may utilize, such as your ecommerce platform"^[3]. Credit for the honesty. Most merchants still miss it.

Do the arithmetic. £500,000 of subscription revenue on Recharge Starter costs you $1,188 a year in platform fees plus roughly £7,450 in percentage fees plus the 19c per transaction, and all of that sits on top of the 1.5% to 2.9% your gateway already takes. Call it 3% to 4.5% all in on recurring revenue.

The Swell anomaly, and what it tells you

Swell includes subscriptions on its $29 Starter plan. Not as an add-on. As a core capability, on every plan including the cheapest^[5].

That is a genuinely different commercial posture to the rest of the field, and it's worth understanding why rather than just noting it. Swell's plans are gated on annual sales volume rather than on features: $50K on Starter, $250K on Basic, $1M on Standard, $5M on Unlimited, with overage percentages of 2%, 1.5%, 1% and 0.4% respectively^[5].

So the money still scales with your revenue. It just doesn't scale specifically with your subscription revenue, and you don't pay a separate vendor to touch your billing. For a store where subscriptions are most of the business, that's structurally cheaper. For a store doing occasional subscriptions on a large one-off catalogue, it's probably not.

The trade-off nobody mentions: Swell is a much smaller ecosystem. Fewer apps, fewer agencies, fewer people who've hit your problem before. That's a real cost and it's not on a pricing page.

Medusa has no subscriptions module, and says so

This one deserves its own section because it's the clearest illustration of what "composable" means in practice.

Medusa's own documentation is direct about it. There is no built-in subscriptions module. What exists is a recipe: instructions for building your own subscription module, defining the data models, linking them to orders and products, then implementing either custom renewal logic or a Stripe subscription provider^[6].

The recipe lists what you build yourself^[6]:

• A Subscription Module with data models for interval and period
• Module links connecting subscriptions to orders and products
• A workflow that completes checkout and creates the subscription
• An API route to run it
• A scheduled job checking daily for renewals due
• A second scheduled job checking daily for expiries
• Admin widgets or UI routes so somebody can actually manage them
• Storefront changes for the customer to manage their own

And a specific detail worth knowing: Medusa's standard Stripe provider does not support subscriptions, so the Stripe route also means writing a custom module provider^[6].

The wider Medusa documentation is worth reading before you commit, because the module and workflow model is the thing you will live inside^[20].

I'm not knocking Medusa for this. Being explicit about what you don't ship is more useful than a marketing page claiming subscription support that turns out to mean "you can integrate one". But price it honestly. That list is several weeks of engineering, plus the ongoing maintenance of billing code, which is the least forgiving code in a commerce stack.

Does this actually impact you?

Three questions, and if the first answer is no you can stop.

Is the product genuinely consumable on a predictable cycle? Coffee, supplements, razor blades, pet food, contact lenses. If people run out on a schedule, subscriptions work. If they don't, you're offering a discount for a commitment nobody wants.

Is your repeat purchase rate already decent without it? Subscriptions accelerate an existing habit. They don't create one. If your 90-day repeat rate is 6%, a subscribe-and-save badge won't move it, and you'll spend three months building billing infrastructure to find that out.

Can you take a support call about a failed payment? Because you will. Cards expire, banks decline, addresses change. Subscription businesses live or die on dunning and nobody budgets for it.

What we found reading the vendors' own review pages

We went through the public review data on the first-party subscription products, because it's the one place vendors can't edit the copy.

Shopify Subscriptions: 3.7 stars, 711 reviews, 20% of them one star^[1]. The recurring complaints are not about billing reliability, which is what you'd expect. They're about behaviour at the edges: variant changes silently converting a one-time purchase into a subscription, and customers cancelling immediately after taking the subscribe-and-save discount^[1].

That second one is the finding worth generalising. The most common subscription failure isn't technical. It's that the discount is claimable without the commitment. If subscribe-and-save gives 15% off order one and cancellation is one click, you've built a coupon with extra steps.

The design fix is boring and unpopular: put the discount on the second delivery, not the first. You'll convert fewer subscribers and keep more of them, and your revenue per subscriber goes up. Most merchants won't do this because the first-order conversion number is the one on the dashboard.

Second thing we found: none of the platforms ship dunning worth the name in their free tier. Retry logic, pre-dunning notification before the card is charged, card-updater services that refresh expiring cards through the networks. Those are what separate a subscription business that compounds from one that leaks 5% a month, and they're consistently what you're actually buying when you pay Recharge $99^[3].

The regulation nobody mentions: UK subscription rules

The Digital Markets, Competition and Consumers Act 2024 legislated a whole new regime for subscription contracts in the UK, and it's considerably more prescriptive than what came before^[8]. The obligations it sets out:

• Key pre-contract information presented together, clearly and separately, before the contract is entered into^[8].
• Reminder notices before renewal payments. For contracts with no concessionary period, every six months^[8].
• Cooling-off rights of 14 days after initial entry and on renewal^[8].
• Straightforward exit. Cancellation without steps that aren't reasonably necessary, with an online route for contracts entered into online^[8].
• Acknowledgement notices and refunds of overpayment within 24 hours for online notification^[8].

Commencement of these provisions is staged, so check the current position before you plan around a date. The direction is not in doubt. The US moved the same way with the FTC's negative option rulemaking^[9], and the EU already gives consumers a 14-day withdrawal right on distance contracts, which we covered in the right of withdrawal piece.

Practical read: build the cancel-online flow now. If your cancellation route is an email to support, that's the thing to fix first, ahead of anything to do with growth. It's also, incidentally, what reduces chargebacks.

If you are on Shopify

How subscriptions actually work here

Shopify doesn't do recurring billing itself. It exposes three API surfaces that a subscription app uses: Selling Plans to define the offer, Subscription Contracts to hold the agreement between merchant and customer, and Customer Payment Methods to store the credential for future charges^[10]. The contract is the object that matters, and it's what you'd have to migrate if you changed vendor^[11].

Free versus paid, honestly

Shopify Subscriptions is free and adequate for a simple subscribe-and-save on a small catalogue^[1]. Recharge at $99 plus 1.49% buys you dunning, bundles, customer portal depth and migration support^[3].

Bold is the other long-standing option in this tier and worth a quote for comparison^[17].

The break-even is roughly: if subscriptions are under about 10% of revenue and the offer is simple, start free. If subscriptions are the business, the percentage fee is the price of not building a billing team, and it's probably worth it until it very obviously isn't, which is somewhere north of £2M of subscription revenue where 1.49% is £30,000 a year.

The variant trap

Worth testing on your own store specifically. Merchants report the purchase option flipping from one-time to subscribe when a customer changes variant^[1]. Go and change a size on a subscription product and watch what the radio button does.

Not on Shopify? The other platforms

WooCommerce

WooCommerce Subscriptions is £213 a year with no per-transaction cut^[4], which makes it the cheapest serious option in this comparison by a distance once you pass modest volume. At £500,000 of subscription revenue you'd pay £213 rather than roughly £8,600 on Recharge Starter.

What you're taking on instead: gateway compatibility is your problem. Not every Woo payment gateway supports tokenised recurring payments properly, and SCA handling on renewals is where the badly-maintained ones fail. Check your gateway supports the recurring-transaction exemption path before you sell a single subscription^[12].

Magento and Adobe Commerce

No native subscriptions. You're buying an extension or building against the order management layer^[13]. The REST API is the integration surface you would build against^[19]. For a mid-market Magento merchant this usually ends up being a custom build integrated with an ERP, which is the right answer when subscriptions are genuinely operationally complex and the wrong answer when you just want subscribe-and-save.

BigCommerce

No native recurring billing either. Third-party app or a Stripe Billing integration against the Orders API^[14]. BigCommerce's embedded payment provider list matters here, because the subscription app you choose has to work with the gateway you're not being charged extra for^[7].

Headless, Medusa and custom

If you're building on Stripe Billing directly, the primitives are good and well documented: subscriptions, invoices, retries, and the customer portal^[15]. The work you still own is reconciling Stripe's subscription state with your commerce platform's order records, which sounds trivial and is where most custom implementations rot.

Stripe's own build guide is the reference implementation and it is genuinely good^[18].

One specific thing: store the payment method with setup_future_usage at the first payment so the initial authentication carries forward^[16]. Getting this wrong means every renewal triggers a fresh authentication that nobody is present to complete.

The subscription decision table

If this is true	Route	Why
Testing the idea, under £10k/mo subscription revenue	Shopify Subscriptions, free	No commitment, adequate for simple offers[1]
Subscriptions are 30%+ of revenue on Shopify	Recharge or equivalent	Dunning and portal depth are what you're buying[3]
Subscription revenue over ~£2M on Shopify	Model moving off the percentage	1.49% becomes a headcount[3]
On WooCommerce, competent gateway	WooCommerce Subscriptions	£213/yr flat, no revenue share[4]
Subscriptions are the entire business, greenfield	Swell	Bundled from $29, no separate vendor[5]
You have engineers and unusual billing logic	Medusa or Stripe Billing direct	Full control, real build cost[6]
Your repeat rate is under 10% at 90 days	Don't	Fix retention first

What agents change about recurring orders

Less than the hype suggests, and in one specific way more.

The less: nobody's AI assistant is going to spontaneously subscribe you to coffee. Recurring commitments are exactly the decision people don't delegate.

The more: agents are extremely good at cancelling things. A shopper who asks an assistant to find and cancel their unused subscriptions will find yours, and the assistant will follow whatever cancellation path you've published. Which means the DMCC requirement for a straightforward online exit^[8] is about to be enforced by software rather than by regulators.

Businesses built on friction-to-cancel are going to have a bad few years. That's not a prediction I'd have made confidently two years ago and I think it's now the safest one in this article.

What to do this week

• Calculate your true subscription cost. Platform fee plus percentage plus per-transaction plus gateway. One number, as a percentage of subscription revenue.
• Check your 90-day repeat rate before building anything. If it's low, this is the wrong project.
• Test the variant switch on your own subscription products^[1].
• Find out what happens on a failed renewal. How many retries, over how many days, with what customer communication. If you don't know, you have no dunning.
• Try to cancel your own subscription as a customer. Time it. If it takes more than a minute or requires an email, fix it^[8].
• Move the subscribe-and-save discount to the second delivery and watch what happens to retention.
• Confirm your gateway handles the recurring-transaction SCA exemption^[12].

The takeaway

The cost spread across platforms on the same £500,000 of subscription revenue runs from £213 a year to roughly £8,600, and the expensive option is often the correct one because what you're buying is dunning rather than billing^[3][4]. Swell bundling it at $29 is the outlier worth knowing about^[5]. Medusa not shipping it at all, and documenting exactly what you'd have to build, is the most useful vendor honesty in this whole comparison^[6].

I got the discount design wrong for a long time. Front-loading the saving onto the first order made the launch metrics look great and produced a cohort that churned at the second delivery. That took two quarters to notice because the acquisition number kept going up.

How long does it take to cancel a subscription on your store? Go and time it.

Sources

1. Shopify, "Shopify Subscriptions app listing", Shopify App Store (711 reviews, 3.7 average). Accessed 22 July 2026.
2. Shopify, "Shopify pricing". Accessed 22 July 2026.
3. Recharge, "Recharge pricing". Accessed 22 July 2026.
4. WooCommerce, "WooCommerce Subscriptions". Accessed 22 July 2026.
5. Swell, "Swell pricing". Accessed 22 July 2026.
6. Medusa, "Subscriptions Recipe", Medusa documentation. Accessed 22 July 2026.
7. BigCommerce, "BigCommerce pricing". Accessed 22 July 2026.
8. UK Parliament, Digital Markets, Competition and Consumers Act 2024, Part 4 Chapter 2 (Subscription Contracts), legislation.gov.uk.
9. US Federal Trade Commission, "Negative Option Rule". Accessed 22 July 2026.
10. Shopify, "Subscriptions", shopify.dev. Accessed 22 July 2026.
11. Shopify, "Subscription contracts", shopify.dev. Accessed 22 July 2026.
12. European Commission, Commission Delegated Regulation (EU) 2018/389, Article 14 (recurring transactions), EUR-Lex.
13. Adobe, "Orders, Adobe Commerce Admin Guide". Accessed 22 July 2026.
14. BigCommerce, "Orders". Accessed 22 July 2026.
15. Stripe, "Subscriptions overview", Stripe Billing docs. Accessed 22 July 2026.
16. Stripe, "Save a payment method during payment". Accessed 22 July 2026.
17. Bold Commerce, "Bold Subscriptions". Accessed 22 July 2026.
18. Stripe, "Build a subscriptions integration". Accessed 22 July 2026.
19. Adobe, "Adobe Commerce REST API". Accessed 22 July 2026.
20. Medusa, Medusa documentation. Accessed 22 July 2026.