Content that maximises engagement might actually destroy your subscription business.

Academics, commentators, and many others believe the shift from ad-funded to audience-funded media contributed to increased political polarisation. 

For example, media scholar Andrey Mir wrote in his book Postjournalism: “The ad‑driven media manufactured consent. The reader‑driven media manufacture anger. The former served consumerism. The latter serve polarisation.”

Now we learn that actually the opposite might be true. A new research based on novel methodologies and using real publisher data tackles this big question of our time: Does polarising news content really pay and how?

Value-destroying content: Professor Shunyao Yan of Santa Clara University and Professor Klaus Miller of HEC Paris analysed data from “a major European news Web site” (62,000+ articles, 40 weeks of user-level clickstreams and conversions), used LLMs to score content, and isolated causal effects. 

They found that polarising, and especially emotionally charged, “us-vs-them” content, increased time spent but reduced the likelihood of subscribing, particularly during politically charged moments like elections.

They concluded: “For publishers with a subscription-focused business model, a strategy of leveraging high-arousal, effectively polarising content is value-destroying. While such content may capture fleeting attention, it appears to erode the very foundation of trust and perceived value necessary to convert a reader into a paying subscriber.”

Quantifying the impact on revenue: I took the academics’ findings and applied them to a simple model using subscription and advertising benchmarks from 289 news brands worldwide, which we track at INMA. 

For every 1 million online users active in a month, a median national news brand sees:

• ~500 new digital-only subscription starts.

• ~$75k in total subscription lifetime value created.

• ~$69k in total digital ad revenue generated.

Now introduce polarising content:

• A modest 5%-10% drop in subscription conversion destroys US$4-US8k in subscription LTV.

• A generous 5%-0% uplift in non-subscriber engagement adds only US$2-US$5k in ad revenue.

Even before we talk about churn or brand effects, the economics already breaks. Tiny damage to subscription conversion wipes out large engagement gains.

For news businesses, this research by Professors Yan and Miller suggests a hard truth: Not all engagement is worth it, and some of it is actively value-destroying.

Subscriptions don’t incentivise polarisation. They punish it.

Curious about INMA Subscription Benchmarking Service? Drop me an e-mail and book a demo with a briefing on the latest findings.

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