FT transforms subscriber retention through AI and personalisation

The current media landscape sees news companies grappling with rising acquisition costs, increased competition for readers’ time, and higher-than-ever expectations for personalised, seamless experiences. Coupled with subscription fatigue and a growing tendency for readers to re-evaluate the value they get, the case for doubling down on retention has never been stronger.

While the FT has long focused on improving customer experiences, embedding AI into our core retention journeys marks a significant advancement in our data-driven, customer-centric strategy. It enables us to meet subscribers’ evolving needs by ensuring they are on the right product, at the right price, and at the right time.

The retention strategy evolution

In 2019, the FT introduced a self-serve online cancellation journey to improve both customer retention and user experience. After observing that customers who had a tailored, personal conversation with our customer care team were more likely to stay subscribed, we set out to replicate that effectiveness online.

The journey evolved to include save offers based on business logic, usage insights, and transparent messaging to help customers make informed decisions about their subscription.

The strategy worked. We saw measurable improvements in retention and sentiment, but we knew we could go further by using AI to enhance decision-making at critical moments and deliver greater personalisation at scale.

Instead of manually defining rules based on historic performance, AI allows us to respond dynamically to customer behaviour. It can adapt as patterns shift, which is a crucial capability in a fast-changing media landscape, especially with 2025 set to be a turbulent news year.

Making AI work across the trial lifecycle

Trialists made an ideal starting point for this strategy. As a low-commitment, price-sensitive group, they pose one of the biggest opportunities for long-term retention. Our goal was to convert more trialists into long-term subscribers through two interventions:

1. Proactive intervention before cancellation. Trialists flagged as high-risk in the final week of their trial receive a pre-emptive e-mail offer recommending a more suitable product, payment term, or discount.
2. Reactive intervention at cancellation. When a trialist attempts to cancel online, the model recommends a tailored alternative, such as a different product, payment term, or discount, based on their behaviour and churn likelihood.

To bring this to life, we partnered with AI vendor Vector Labs, whose expertise in model development and implementation was pivotal.

Building, training, and integrating the models into existing journeys was a major collaborative effort, involving teams across marketing, product and tech, data science, business intelligence, data platforms, payments, and more.

Inside the process

The idea was scoped in late 2023, with development beginning in early 2024 and the first experiment launching in October. One challenge was scale. Unlike acquisition journeys, cancellations occur in lower volume, so we had to be strategic in testing.

With such a complex integration, we couldn’t risk cohorts experiencing poor performance, such as long load times or dead ends. Granular tracking and close monitoring ensured the experience remained seamless.

We also applied ethical safeguards throughout, including model constraints, regular performance reviews, and oversight from internal governance teams to help prevent bias and ensure responsible use.

Personalised products, payment terms, and offers are recommended by an AI model and then sent to high-risk trialists before they churn.

Lessons from the first phase

We ran an eight-week controlled test, with half of the trialists exposed to AI-driven offers and the other half continuing with our previous logic. The results were very encouraging:

• 113% increase in save rate for trialists attempting to cancel online.
• 51% uplift in LTV per capita for those entering the cancellation journey.
• 165% improvement in trial conversion rate through proactive e-mail outreach.

Beyond the numbers, the project showed that AI can be embedded into retention workflows, reducing manual rule-setting and enabling more adaptive, data-led strategies.

What’s next

As we conclude phase one of our AI-driven retention strategy, there’s plenty ahead:

• We’re monitoring model performance to ensure continued effectiveness as audience behaviours evolve.
• We’ve launched a redesigned cancellation journey for trialists, with clearer pricing and improved mobile UX, which is already having an uplift on retention.
• We plan to expand the model’s scope by including lower-cost propositions like FT Edit and the digital edition.
• We’re preparing to apply AI strategies to non-trialist segments to support core subscriber retention.

Redesigned onsite cancellation journeys will better articulate the AI-recommended save offers, especially on mobile as a preferred channel.

We see this initiative as the beginning of a wider transformation in how we retain and grow our subscriber base. For news publishers about to commence their AI retention journeys, our advice is simple: start small, test rigorously, and always keep the customer at the centre.

At FT Strategies, we help publishers build retention and revenue models that are grounded in audience insight, rigorous testing and long-term value creation. Whether you’re exploring AI-enabled journeys, churn diagnostics or personalised lifecycle strategies, book a call with our expert team to see how we can create measurable impact for your organisation.