Publishers have a problem: Their paywalls are too steep to scale

By Christian Printzell Halvorsen, CEO, Cxense

If 2018 showed us anything, it’s that the Duopoly seems to be unstoppable. Google and Facebook lurched from scandal to scandal, suffering bad press, but sacrificing just a fraction of their market shares. Why? Put simply, they still have more of the granular, behavioral and interest data that advertisers want.

To bolster themselves against this unshakable reality, publishers are doubling down on subscriptions. In fact, 52 percent said building logged-in audiences will be their top business priority in 2019, according to Reuters Institute. Bloomberg, Vanity Fair, Business Insider and others are already working on it, gathering both recurring revenue and the deterministic IDs that capture accurate first-party data.

But the strategy is about to hit a wall. Specifically, a hard (or metered) paywall. Publishers often deploy static paywalls — the kind that drops uniformly for every reader — as a first step toward building a subscriber base. These initially pay dividends by accruing a base of frequent readers and super fans. But publishers soon find that it is trickier to lure occasional and flyby readers with the same approach.

Enter dynamic paywalls. Last year, the New York Times, New York Media, and the Wall Street Journal launched new, data-driven paywalls that court subscribers in a more personalized way. Where static paywalls are essentially one-size fits all, the machine learning behind dynamic paywalls crunches user data before determining how to meter site content and offer discounts in return for memberships.

Variables might include demographic information like geography and salary range, as well as on-site behavior like how many stories they read in a session, what kinds of stories they read and more. The Wall Street Journal, which hit its 3 million subscriber mark last year, did just this: bucketing readers into three groups and serving offers based on their likelihood to subscribe.

Publishers don’t need to have the Journal’s resources or scale to accomplish their subscription goals. Danish publisher Nordjyske Medier increased its subscribers by 50 percent when it used a dynamic paywall to tailor and test subscription messages. Through this method, they soon found that die-hard soccer fans were more likely to subscribe when encouraged to keep up with their team’s every play.

The Journal’s approach is simple and effective, but dynamic paywalls offer myriad possibilities. Publishers can balance ad campaign fulfillment with subscription revenue by gating content based on interest. Those die-hard soccer fans might not have access to the sports section without a subscription, but have free reign of the other sections where they contribute valuable impressions. Further, when a big story breaks, editors who want to make sure it gets the widest reach (and ad execs hungry to meet impressions quotas) can request the paywall be dropped.

Most importantly, publishers can use modelling to predict which customers are likely to churn, and when. The Times and Sunday Times of London found that 96 percent of the readers who entered the subscription process were dropping off. Using propensity modeling (a statistical scorecard used to gauge the likelihood of a consumer’s behavior), they began segmenting readers and personalizing their message. The changes boosted conversion by 190 percent according to recently released figures. The same can be done with users who are at risk of cancelling their subscriptions.

But dynamic paywalls aren’t just about the bottom line. User experience is buoyed when publisher sites respond to an individual’s distinct preferences. After years of social feed algorithms catering to their every whim, readers expectation for personalized content is higher than ever. Those raised expectations open up an opportunity for publishers.

For most of the industry’s history, publishing has been a top-down, mass media experience. Reporters and editors dictated the shape of the world based on their own priorities, and in large part they still do. But the one-to-many dynamic that worked when people were accustomed to being spoken to no longer does. Now, readers no longer want to be just another set of eyes. They want to be a valued customer.

Personalized experiences allow for this. Much like subscription services that curate selections based on a customer’s stated preferences, content personalization can design home pages, app screens, newsletters and advertising messages that speak to an individual’s needs. Publishers that opt-in to personalization move from courting readers to courting customers and, eventually, establishing consumption habits like daily newsletter reading. With tactics like this, publishers have the opportunity to become more relevant than ever.

This service-oriented mindset is just what is needed to steal attention (and advertising money) from the almighty platforms, where algorithms pick the content and “users” aren’t just one of thousands: they’re one of billions. At the end of the day, the Duopoly may be able to reap data from an increasingly skeptical audience, but overcoming the brand relationship between a trusted media outlet committed to personalization and its subscribed readers may be too steep a wall for them to climb.

https://digiday.com/?p=323873

More from Digiday

At the Las Vegas Grand Prix, Mastercard joins a pack of consumer brands flocking to Formula One

For marketers looking to align their brands with F1’s expanded appeal to audiences, the Las Vegas Grand Prix is providing a slip road into the sport.

News publishers may be flocking to Bluesky, but many aren’t leaving X

The Guardian and NPR have left X, but don’t expect a wave of publishers to follow suit. Execs said the platform is still useful for some traffic and engaging with fandoms – despite its toxicity.

Buying with bots: AI search raises the bar for tailored shopping and transparency

AI search platforms like Perplexity and Amazon are adding new ways to shop, but where do the generated recommendations come from?