Social Rocket
Reuters is relying on a dynamic paywall that uses machine learning to decide when users are shown a subscription offer. Unlike a rigid metered paywall, it factors in signals such as user behavior, content performance, conversion probability and the expected ad value of a visitor.
Digiday reported on the model on 16 September 2026, following an appearance by Phil Andrao, General Manager Digital at Reuters, at the Digiday Publishing Summit.
The paywall does not only optimize for subscriptions
The interesting point for marketers and publishers is that Reuters does not treat subscription and advertising revenue as opposites. Subscribers continue to see ads. According to Andrao, they consume more content, visit the website and app more often and are authenticated. This creates additional first-party data potential for the advertising business.
According to Digiday, the machine-learning system estimates, among other things, how much ad revenue Reuters could generate from a user and how likely a subscription conversion is at the same time. Based on these signals, the system decides whether a paywall appears.
The news cycle also plays a role. During periods of especially high demand, the paywall can become stricter because content converts better. On quieter days it is relaxed.
Reuters is sticking with one global price
Reuters officially launched its digital subscriptions in October 2024. The offer initially started in Canada and was then expanded to further markets.
According to Digiday, the subscription now costs 4 US dollars per month worldwide and is available in 80 countries. Reuters reportedly deliberately avoids introductory prices and discount promotions.
Andrao did not give a current figure for the total number of subscribers. Reuters President Paul Bascobert had spoken in June of more than 2025 subscribers 100.000.
Why this matters for advertisers
For publishers, the paywall is not just a pricing problem. Authenticated, recurring users can be more valuable to advertisers than a very large but fleeting audience. Andrao therefore argues that engagement can matter more than raw monthly unique figures.
This fits a broader trend in the advertising market: first-party data is gaining importance, while third-party signals are becoming less reliable and AI systems are taking over additional parts of the customer journey. The IAB itself currently names publishers with first-party data as an area where advertisers are increasing their focus.
In parallel, technical measurement is becoming harder. Our article on Bot traffic, retargeting and AI agents shows how automated access can distort analytics and campaign logic. Google is also working on new CrUX metrics for the real ad load on websites. And as AI answers grow in importance, the measurement of AEO and AI visibility KPIs is becoming more relevant too.
Reuters blocks bots more selectively
Bot management is therefore part of the strategy as well. According to Digiday, Reuters has been using a "block by default" approach since May 2026 and only allows selected crawlers. Google is currently still permitted to crawl.
The decision is economically motivated: if a bot delivers neither audience nor advertising or subscription value while using content for its own products, the incentive for Reuters to allow it decreases.
For media houses, this creates an increasingly integrated model of subscription optimisation, advertising value, first-party data and bot control. The paywall is no longer just an access barrier but part of revenue management.