What happens after a reader hits your paywall? Most leave the Web site. Some desperately reload the page or navigate to the home page. Many learned how to successfully bypass the paywall.

A team of researchers from Georgia Tech and the University of Notre Dame in the United States studied how readers respond to paywalls and discussed their results with INMA.

They used clickstream data from a large North American news publisher with a client-side, metered paywall. The clickstream data allowed the academics to observe actual behaviours rather than relying on surveys and self-reported reader journeys.

Buyers, leavers, and dreamers

The good news is that paywalls are confirmed effective even if they don’t lead to immediate conversions.

In line with what we see in the INMA Subscription Benchmarks, the academics found only 0.21% paywall stops resulted in subscriptions. 

Still, this conversion rate was 50 times higher than the rate among visitors who hadn’t seen the paywall at all. “Almost no one comes to the Web site to subscribe without first seeing the paywall,” Professor Adithya Pattabhiramaiah from Georgia Tech said.

So what did readers do if not subscribing? 

• In 57% instances, they simply left the Web site. 

• 21% of the time, readers stayed and read non-premium content, such as the home page or section pages. 

• Readers also probed the meter 12% of the time, trying to access other articles. Unfortunately for them, this also triggered the paywall. 

• Some (10%) responded by reloading the page, perhaps in a hope the paywall goes away.

Paywall evaders

The bad news is the researchers saw readers successfully bypassing the paywall after 11% of stops.

How did they figure it out? Hard to say about evaders, but the researchers pioneered a novel method of measuring the evasion behaviour. 

They tracked cases where a user viewed the same article twice in quick succession — once triggering the paywall and once not — from what appeared to be the same user (based on IP address, user agent, and other signals). 

This “article-as-anchor” method shed light on paywall evasion even when cookies were refreshed or deleted.

“Tracking browser cookies doesn’t work for detecting paywall circumvention because many circumvention methods generate a new cookie,” explained Professor Eric Overby from Georgia Tech. 

For example, the most common evasion method was switching to the browser’s private mode (82% of instances). Other ways, such as switching a browser or disabling JavaScript, were used too but less often. 

“Our key insight is that readers who circumvent are often doing so to access the article that triggered the paywall.”

How to use this

Both responses to the paywall — subscribing or bypassing it — reflect reader engagement. 

The academics analysed what factors predicted reader responses. Interestingly, both subscribers and evaders tended to visit directly rather than from social media or search. They also accessed opinion articles more than other content.

In other words, the evaders love your content but instead of paying money, they pay with effort.

“Publishers can use our method to assess which readers are circumventing the paywall,” advised Professor Vamsi Kanuri of Notre Dame. “This can help better understand those readers and, for example, target them with promotions.”

Interested in learning more? Read the full research paper. E-mail Professor Eric Overby at eric.overby@scheller.gatech.edu.

Greg’s Readers First newsletter is a public face of a revenue and media subscriptions initiative by INMA, outlined here. INMA members can subscribe here.

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