Think about the last time you tried to cancel something. A gym, your broadband, any subscription. Five screens of “are you sure?”, a short survey and a special offer at the end, and sometimes you just give up and stay.
Most publishers run a version of that flow, and it works, up to a point. The problem is, some readers only stay out of habit and are still paying because leaving is a hassle.
That flow is about to get much shorter.
Hand the same job to a reader’s AI assistant and it’s done in seconds. It doesn’t get tired, it doesn’t get distracted by the survey, and it has no reason to feel bad about leaving. So the hassle that has been keeping some readers subscribed is going to stop working - and probably pretty quickly.
On the other hand, while hassle has kept some readers in, it’s also kept a lot more people out.
- The person who wants one story rather than a subscription.
- The person who'd pay for three months of local coverage but not a year of everything.
- The person who'd happily pay two dollars but won't fill in a form to do it.
The same technology that makes it easier than ever to stop paying for news could make it easier than ever to start.
What a good answer looks like
Say Tom pays $12.99 a month for a bundle, and his assistant has been asked to cancel it because he only uses the recipes and the puzzles. The publisher’s first response is the standard one, 10% off everything for three months, and the assistant turns it down flat, because Tom’s budget is $7.
So the publisher offers something different: recipes and puzzles, ad-free, for $8.99. Still too much.
Recipes and puzzles with ads, then, at $6.99 a month or $69 a year. The assistant takes the annual plan.
The whole exchange takes seconds. The package changed, cut down to the two things Tom actually uses, inside a price floor the publisher had set in advance. Tom keeps the part of the product he values, and the publisher keeps Tom for another twelve months.
The relationship doesn’t end at cancel
Most news publishers have about twice as many former subscribers as current ones, according to INMA’s analysis of its subscription benchmarks last year. The median brand has lost more digital subscriptions since 2019 than it has active today, and almost none of those people were offered anything other than the product they’d just decided to stop paying for.
That’s because people don’t subscribe in a straight line. They read for free, trial, subscribe, pause, let a card fail, come back for a different bundle, and before long many of them will be doing all of that through an assistant.
So readers’ agents will cancel, and some of those cancellations are fine. The ones worth having a plan for are the readers who’d stay for a different bundle, discount or plan, if anyone offered it in time.
What it takes to answer in seconds
To answer in seconds, you need three things.
- A product you can take apart and put back together in different ways, so there’s something to offer other than a one-size-fits-all subscription.
- Rules that say how far an offer can bend for each kind of reader, including a floor you won't go below.
- A system that can put the two together at the moment the request arrives, rather than a week later in a win-back email.
Your team sets the goals and the guardrails, and before long an assistant on your side does the negotiating. That’s what agentic monetization makes possible, and it’s worth having the groundwork in place now, so that when the request arrives all three parts are ready.
MonetizationOS is the layer that decides who gets access to what, for every visitor, human or machine, along with the offers and rules behind those decisions. If you’d like to talk through what it would take to get your product and your rules ready for that, talk to our Head of Revenue Scott Howland, [email protected].
