Go to Member Center

Reuters Dynamic Paywall Raises a Personalization Question

ftc paywalls personalization reuters subscription strategy Sep 17, 2026

Reuters says its machine-learning paywall is growing subscriptions while supporting advertising revenue, even as estimated site traffic falls. The way it decides when individual readers encounter the paywall raises another question: What does personalization mean when the subscription price itself stays the same?

Reuters is using machine learning to decide when individual readers encounter its subscription paywall, weighing the potential advertising value of a reader against the likelihood that the person will subscribe.

Phil Andraos, general manager for digital at Reuters, told the Digiday Publishing Summit this week that the dynamic paywall has been fundamental to growing the company's subscriber base. He also said Reuters' advertising business is performing better than it was before the paywall launched earlier this year. Subscribers continue to see ads, consume more content than nonsubscribers and return more frequently. 

Those claims come against a notable traffic decline. Similarweb estimates Reuters site traffic fell 19% year over year in August 2026. Reuters has not attributed that decline to the paywall. The company has also reduced bot traffic, while Andraos said the traffic Reuters can monetize has remained unchanged.

Reuters has not disclosed its current subscriber count or quantified the improvement in advertising performance. Andraos declined to provide an updated subscriber number. Reuters President Paul Bascobert said in June 2025 that the company had more than 100,000 subscribers. 

Dynamic paywalls are not new to publishing. Publishers including The Washington Post, Los Angeles Times, The Wall Street Journal, Barron's and Quartz were already using dynamic models several years ago to determine which readers were most likely to subscribe and which were more valuable as advertising-supported visitors.

How Reuters Decides When to Show the Paywall

Reuters' machine-learning system considers signals including user behavior and how well particular content has converted readers. It estimates how much advertising revenue Reuters could generate from a user, then considers that person's likelihood of subscribing before deciding whether to display the paywall. 

The system also responds to the news cycle. Reuters tightens the paywall when news activity increases and readers are more likely to subscribe, then loosens it during slower periods when conversion propensity falls.

The subscription price does not change. Reuters charges a flat $4 per month across 80 countries and does not use introductory or promotional pricing. 

The price can therefore remain uniform even as the point at which a reader encounters it changes.

The Same Price, But What Is Being Personalized?

The Federal Trade Commission is currently seeking comments on a proposed enforcement policy statement addressing personalized pricing. The FTC describes personalized pricing as using personal data to set prices according to the amount a company believes an individual consumer is willing to spend. Public comments are due September 25

Reuters describes a different kind of decision. Its subscription price remains $4 per month, while its machine-learning system uses behavioral and performance signals to decide whether to show a particular reader the paywall. Reuters has not publicly detailed all of the inputs used by that system. 

Insider Take

Reuters is not simply trying to maximize subscriptions. Its system is weighing whether a reader is more valuable at that moment as an advertising-supported visitor or as a potential subscriber.

That makes the reported 19% traffic decline important context, but not enough on its own to evaluate the strategy. Reuters says its advertising business is performing better, subscriptions are growing and monetizable traffic has remained steady. But the company has not disclosed current subscriber numbers or quantified those improvements, so the available data does not tell us how large the gains actually are. 

For operators, there are two issues worth watching. The first is measurement. Reuters is arguing that raw traffic matters less than the economic value and engagement of the audience that remains.

The second is personalization. Reuters keeps the subscription price at $4, but its algorithm helps determine when an individual reader encounters that price. At the same time, the FTC is asking businesses and consumers to consider a different use of personalization: changing the price itself based on information about the consumer. 

As subscription decisioning gets more sophisticated, the question for operators is not only whether two customers see the same price. It is what else the algorithm is changing around that price.

 

Related Member Resources

Reuters is using a machine-learning paywall to decide when readers encounter a subscription offer while balancing paid growth with the advertising value of its audience. These resources help operators examine two parts of that decision: whether the paid product gives readers enough reason to subscribe, and whether the growth produced by the paywall is creating lasting business value.

  • Paid Content Strategy: What Makes Readers Willing to Subscribe?

    This research looks at what readers say makes digital content worth paying for and where perceived value can break down. For a publisher using a dynamic paywall, it helps teams examine whether the timing of the paywall supports conversion and whether the underlying paid offer gives readers a strong enough reason to begin an ongoing subscription.

  • Subscription Growth Quality Framework: Beyond Subscriber Adds (Includes Scorecard)

    This framework and scorecard helps operators look beyond subscriber additions to retention, engagement, revenue quality and long-term business value. That is useful when subscription growth is occurring alongside changes in traffic or advertising economics because it helps teams assess what kind of growth the paywall is actually creating.

Sources