This week the UK government announced that from January 2027 it will, in effect, be illegal to run a subscription business the way most subscription businesses are actually run.

The BBC called it a ban on "subscription traps", brought forward three months from its original spring 2027 start as part of the Prime Minister's cost-of-living push. The legislation behind it - the Digital Markets, Competition and Consumers Act 2024, being switched on in stages - calls it the new subscription contracts regime. I call it the moment the UK subscription economy finds out how much of its revenue was ever real.

I read the announcement the way a reformed arsonist reads the new fire code: nodding along, faintly nostalgic, and in absolutely no position to be outraged.

Some context for newer readers. I spent 22 years building subscription businesses, most of them in online dating. White Label Dating, the platform I started in 2003, powered around 25,000 dating sites for 1,700 partners at its peak - 45 million members, roughly $50m a year in revenue, half a billion dollars billed over its lifetime, bootstrapped from an office in Windsor. I have spent more hours of my life thinking about renewal billing than any human reasonably should, and I have the war wounds - and the CMA paperwork - to prove it.

So believe me when I say: January is a much bigger deal than the coverage suggests.

What's actually changing

The new rules, confirmed in April's consultation response and now accelerated, are worth reading in full if you run anything with a renew button. The short version:

  1. You must tell people what they're signing up for, before they sign up. The total cost, what happens when the trial ends, when it renews, how to leave - presented clearly and separately, not buried in clause 14.3 of the terms
  2. You must remind people you're about to bill them. Roughly every six months for monthly products, before the first paid payment after any free or discounted trial, and earlier still before anything that locks in twelve months or more. On a durable medium - email, SMS, even WhatsApp counts. A vanishing in-app notification does not.
  3. Cancelling must be straightforward. If they joined online, they leave online. "Email us to cancel" is unlikely to cut it. "Cancel your direct debit with your bank" definitely doesn't. You can make a save offer; you cannot build a labyrinth.
  4. Cooling-off rights now attach to renewals, not just sign-ups. Fourteen days after a trial converts to paid, and after any renewal that commits someone for twelve months or more - with proportionate refunds, paid within fourteen days. The streaming industry lobbied hard for a "binge and cancel" exemption. It lost. Get your notices wrong, incidentally, and the cooling-off window can stretch to a year.

(Charitable museum and heritage memberships are exempt. The National Trust may stand down.)

None of this is theoretical, because the enforcement machinery is already running. Since April 2025 the CMA can decide consumer-law breaches itself - no court required - and fine up to 10% of global turnover. It has already drawn blood: a £4.2m penalty against the AA for drip pricing this April, £900,000 from StubHub, £720,000 from Marks Electrical, an open file on Gold's Gym's joining fees, and a reported probe into Adobe's "annual contract, billed monthly" termination fees. The regulator that spent decades writing stern letters has developed a taste for percentages.

And if your compliance plan was "wait for the final guidance" - the guidance isn't finished, and the deadline moved anyway. Welcome to the new pace.

The £1.6bn that was never really yours

Here's why this is existential rather than administrative.

The UK holds about 155 million active subscriptions worth roughly £26bn a year.

By the government's own count, nearly 10 million of them are unwanted - about £1.6bn a year paid for things people don't use, don't remember, or have actively tried to escape. Three and a half million people have rolled off a free trial into a paid plan without clocking it. Citizens Advice puts spending on unused subscriptions at £688m in a single year. National Trading Standards reckons 4.7 million adults are currently paying for a subscription they didn't even know they had.

Inside the industry we have a name for these customers, and it isn't flattering: zombies. The gym member whose last visit was February. The streaming account still paying for a service whose password nobody remembers. The dating profile - I'll come back to dating - quietly renewing for someone who met their partner in 2023.

Zombie revenue is wonderful. No support costs, no usage costs, no complaints. Pure margin. It is also - and I say this with love - not revenue you earned. It's revenue you found. And the Act is a controlled detonation underneath it.

My favourite detail in the whole saga is buried in the legislation's impact assessment: officials costed the revenue businesses will forgo once exits are easy and reminders are mandatory, and it comes to well over £100m a year. Parliament has, in effect, pre-booked your churn. When the Treasury can forecast the hole in your P&L more honestly than your own board deck does, it is time for a different board deck.

Inertia is not loyalty. It just books the same. Until January.

Dating built the machine. I held the spanner.

Every industry has dabbled in dark patterns, but online dating industrialised them, because dating has a uniquely awkward engine: success means the customer leaves. When your product's entire job is to make itself unnecessary, the temptation to keep billing the distracted is enormous. Add white-label economics - thousands of partner brands, all wanting revenue maximised this month - and you get the industry regulators have been making examples of for a decade.

The rap sheet is instructive.

In 2014, the FTC's first-ever action against an online dating service hit JDI Dating - an English white-label operator, as it happens - over "Virtual Cupids": fake profiles that messaged free members with romantic interest until they upgraded, plus auto-renewals nobody had knowingly agreed to. It cost them $616,165 and their business model.

In 2016, Ashley Madison settled with the FTC over its "engager" bots.

In 2019 the FTC sued Match Group, alleging Match.com had shown non-subscribers "someone is interested" ads generated by accounts Match itself had already flagged as likely scams - nearly 500,000 subscriptions were bought within 24 hours of such ads, per the complaint - and made cancelling deliberately miserable. Match's own internal documents described its cancellation flow as "hard to find, tedious, and confusing." Say what you like - as product documentation, it's accurate. The case settled a year ago this week for $14m and a permanent order requiring simple cancellation (Match admitted no wrongdoing).

And in 2017–18 the CMA swept UK online dating, and my company was in the front row. We gave formal undertakings: clearer warnings before longer subscriptions auto-renewed, cleaner cancellation and pro-rata refunds, more transparent disclosure of how our shared network operated. Fourteen other dating businesses got warning letters. I'm not going to relitigate that era here - I've written elsewhere about what I got right and wrong in that business. But I will tell you what it looked like from the inside, because it is exactly the tension every subscription operator in Britain is about to feel.

Our partners - who kept the majority of net revenue - wanted every billing lever pulled: every trial converting, every renewal defaulting, every price point tested. Our merchant accounts, meanwhile, needed to still exist the following year. Those two facts were at war every single day. Because long before the CMA had teeth, subscription businesses answered to two regulators nobody elected: Visa and Mastercard. Cross roughly 1% chargebacks and you didn't get a consultation paper and a transition period - you got a warning from your acquirer and a short window to fix it before you lost the ability to take payments at all. We lost multiple payments processors over the years for this exact reason.

Nevertheless, we then ran below chargeback thresholds for years, and it wasn't sainthood. It was survival. Aggressive billing is a loan against your own payment processing, and the interest rate is fatal.

Which is why I find the doom-mongering about the new regime slightly hilarious.

The discipline it demands - bill people who know they're being billed, for things they actually want - was always the only durable strategy. The card schemes enforced it with a blunt instrument. The CMA will now enforce it with a scalpel and a percentage of global turnover.

For today's white-label and network operators, the maths gets harsher still. The platform is typically the merchant of record, which means the platform wears the consequences of every partner brand's clever ideas. One aggressive partner used to be able to poison your chargeback ratio; from January they can poison your regulatory standing too, with your global turnover as the denominator.

If you run a network and your compliance model is "partners agree to our terms", the next few months are going to be an education. Build the reminder rails, the exit flows and the refund logic centrally - ship compliance as a feature of the platform - or spend 2027 discovering which of your partners enjoys A/B-testing regulatory liability with your merchant account.

The market already ran the test

Here's what the regulation debate keeps missing: customers moved first.

Match Group's paying users have fallen from an average of 15.6 million in 2023 to 13.8 million by the end of 2025. Tinder's payers dropped from 10 million to 8.8 million in two years - while revenue per payer rose every year.

Fewer people, squeezed harder: the business model of a nightclub in decline.

Bumble's 2025: revenue down 10%, paying users down 11.5%, a third of the workforce gone - and now Bumble apparently facing a desperate fire sale.

In the UK specifically, Ofcom found Tinder had shed around 600,000 users since 2023.

Meanwhile, the one Match brand with a genuine outcome promise - Hinge, "designed to be deleted" - grew payers 17% and revenue 26% at the last count. Sit with that contrast for a moment. The apps monetising frustration are shrinking. The app promising to make itself unnecessary is compounding. The market is already running the manual-payment test. January just makes it statutory.

And note which mechanics dating now leans on to manage decline: heavily discounted intro offers, weekly passes that quietly roll into monthly plans, in-app cancellation flows of varying honesty. That is precisely the machinery the UK regime wraps in reminder notices, cooling-off rights and proportionate refunds. Dating won't be the named target of this law. It will absolutely be one of its case studies.

America, incidentally, ran the same experiment with a different script. The FTC finalised its own "click-to-cancel" rule in 2024, then watched a court vacate it on procedural grounds six days before it took effect - and then extracted a $2.5bn settlement from Amazon over Prime's enrolment and cancellation flows under the old law anyway. Plus $14m from Match, $7.5m from Chegg, and live cases against LA Fitness and Uber. Thirty-odd US states now have their own auto-renewal laws. The lesson from America is not "regulation retreats if you lobby hard enough". The lesson is: the fine arrives either way. Only the paperwork changes.

The standing order test

Which brings me to the only subscription metric I have ever fully trusted.

Forget NPS. Set aside your churn dashboard for a moment - it's mostly measuring how good your maze is. Ask one question:

If every customer had to log in on the 1st of the month and manually press a button marked "Pay", how much of your revenue survives?

Call it the standing order test: would they pay you the way you pay for things you've actively chosen - or are you a direct debit they've stopped noticing?

Whatever number you just arrived at, that's your actual business. That's the revenue attached to a product people would walk across the room to keep. Everything above it is an interest-free loan from inattentive customers, and the Act has just issued the repayment schedule.

The businesses that pass this test should be celebrating this law.

Mandatory reminders become marketing: here's what you did with us this month, here's what's coming next. Easy exit becomes a conversion weapon - Citizens Advice found 54% of people have decided against a subscription purely for fear of the cancellation fight; imagine the acquisition uplift when leaving is visibly painless. A renewal cooling-off period becomes a confidence signal, the way a generous returns policy is. Netflix survives manual payment. The boutique gym with a waiting list survives it. Spotify probably survives it twice.

The businesses that fail the test have five months to change the answer. Not by improving the maze - the maze is being demolished by statute - but by making the product worth an active yes, twelve times a year.

What I'd do before January (if it were my P&L)

  1. Run the zombie audit now. Usage-weight your revenue. A customer who hasn't touched the product in 90 days isn't "retained" - they're pre-churned, with a delay on the paperwork. Size that cohort honestly and model 2027 without it. If the number is survivable, invest. If it isn't, you needed to know yesterday.
  2. Build the reminder rails as product, not compliance. The renewal reminder is about to become the most-read message you send. Ship value in it - usage recaps, what's next, what they'd lose - and a mandated touchpoint becomes a retention asset. Send a bare "you will be charged £12.99" and you've built a statutory churn button.
  3. Rebuild cancellation like you're proud of it. One honest save offer, then a clean exit, then a warm door back. Measure the returners; they'll teach you more than the leavers. Amazon's settlement set the global bar - people leave by the same method they joined.
  4. Reprice the trial funnel. A trial that converts into a 14-day cooling-off with proportionate refunds is a different economic object from a trial that converts into silence. Intro offers now have to convert on product love, not amnesia. Budget accordingly.
  5. Watch chargebacks like it's 2007. Your chargeback ratio has always been the earliest indicator of billing your customers don't accept. It now predicts regulators as well as processors.
  6. Reweight the board pack. In: active-use percentage, reminder engagement, time-to-cancel, refund latency, win-back rate. Out: any slide celebrating revenue from customers you haven't seen since the spring.

The part where I sell you something

This is normally where a post like this says "time will tell". Time will not tell. The date is in the diary, the CMA is warmed up, and roughly £1.6bn of UK subscription revenue is about to be asked - politely, legally - whether it would like to continue existing.

I spent 22 years on the operator's side of this: building a $50m-a-year subscription business, keeping it processable, giving undertakings when we got it wrong, and learning - expensively - that the only recurring revenue worth having is the kind your customers would choose out loud.

These days I advise founders, operators and investors on exactly this territory: pricing architecture, retention that doesn't rely on amnesia, payments and billing strategy, and, increasingly, helping acquirers work out how much of a target's "recurring" revenue would survive the standing order test. (There's a book, Subscribe & Conquer, if you want the long version.)

If any part of this post made you think of a specific line on your own P&L, that instinct is correct - and it is considerably cheaper to act on in August than in January. I'm at rosswilliams.com. Let's talk - while your customers are still asleep.