Research shows 39% of national media outlets in 33 countries charged for online news in early spring 2020, according to my testing of 569 news Web sites with the highest declared usage.

• In February and March, together with my associates, we tested the national news sites of brands listed in the Reuters Institute’s Digital News Report as the most read, watched, and listened in 33 countries in the world.
• We accessed each Web site repeatedly over a week, trying to be stopped by the paywalls and searched for paid value propositions. We used a VPN service to appear local users to the sites.
• We identified 224 sites with a paywall or a paid news product, such as the e-replica of a printed edition or news video streaming.
• In a similar study for the Oxford’s Reuters Institute in 2018, in a smaller sample of 397 brands, I saw 32% charging for online news. 

We found 47% of brands charging for online news seemed to follow the freemium model, in which some content is free to all users and some premium content is accessible only to subscribers.

• 18% of brands had a free Web site but charged for access to e-replica of their print products. This replica-only model has decreased significantly in the last two years; in 2018, 32% of brands featured it.
• 12% of brands appeared to have a meter model, in which readers can read a limited number of stories for free.
• 12% of brands did not have a paywall but asked for donations or offered membership. This model may have increased significantly in popularity, as two years ago only 1% of brands in my sample followed it.
• 6% of brands featured a hard paywall, a slight decrease from 8% in 2018.
• 5% of brands seemed to have a hybrid model, a slight increase from 3% in 2018. We observed combinations of a freemium and a meter, and the paywalls kicking-off based on the reader behaviours and other attributes. 

Our study has a number of limitations. We selected the sites based on Digital News Report, an online survey of news consumers’ preferences that ranks sites based on declared usage, and not actual traffic. We tested only nationwide brands’ sites, as the local and regional news is featured in the survey as a category and not the individual brands. The results of our tests, although providing quite accurate results for the brands we know more of, were not validated by the tested publishers. The growing use of dynamic or adaptive paywalls made it difficult to trigger paywalls on some sites through our testing, so we may have underreported the share of hybrid models.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .