Publishers face a quandary when setting up a digital paywall that boosts subscription sales but limits pageviews for their paying advertisers. The goal is to maximize both sources of revenue, a
process that includes determining what kinds of content are most likely to compel readers into buying a subscription.
Those decisions are going to be increasingly handled by algorithms
that sift through reams of data to predict which stories belong behind a paywall, according to Mather Economics. The consulting firm said it conducted a pilot project
with a major U.S. media brand to measure last-click conversions for subscriber-only articles.
The predictive model relied on natural-language processing and information about the content to
predict the best performance. During the three-month test, the articles recommended by the algorithm generated six conversions per article behind the paywall. That number was three times better than
for recommended articles available to readers for free.
As a result, the publisher saw a 20% boost in subscription start volume and an 8% gain in net digital revenue,
according to Mather.
The test results suggest the process of choosing which articles to put behind a paywall can be automated. That may help to remove some of the bias that
comes with having newsroom employees choose the stories. Journalists want their articles to be as widely read as possible, a goal usually contrary to the limits of a paywall.
Mather expects that makers of content-management systems will add more features to help publishers not only produce and distribute articles more efficiently, but also to automate decision-making
about premium content that’s most likely to convert web visitors into paying customers.