In Search for an Audience-Supported Business Model for Local Newspapers: Findings from Clickstream and Subscriber Data

Abstract

With the decline of advertising revenue, local newspapers must shift their revenue sources from primarily advertising to deriving a larger share from subscription fees. Although existing studies on willingness to pay for online news have examined the determinants of people’s paying intent, this study focuses on what factors drive digital news subscribers’ decision to cancel their subscriptions. We propose a framework that illustrates the sources of negative experiences on news websites—namely the inferior good hypothesis, ad-interference hypothesis, and newsletter intervention hypothesis—and investigate how these elements of negative experiences and news reading behaviors are associated with subscription cancellations. We analyze clickstream data merged with subscriber payment data to investigate these associations from three local news sites. Our findings indicate that regularity of news reading, local news content, using ad blockers, and subscribing to certain newsletters are negatively associated with cancellation. These results provide news organizations with a road map for managing the user experience so that consumers will be more willing to pay for the news.

Keywords:

• Local newspapers
• business model
• subscription
• cancellation
• willingness to pay (WTP)
• user experience
• clickstream data
• survival analysis

View correction statement:

Correction

Acknowledgements

The authors would like to thank the Northwestern Local News Initiative (LNI) and the Spiegel Research Center (SRC) for making the data available for this research. Both LNI and SRC are supported by the Lilly Endowment, Myrta Pulliam Charitable Trust, John Mutz, and several anonymous donors. We are grateful to Tim Franklin, director of LNI, Jonathan Copulsky, director of SRC, and Tom Collinger, former director of SRC and associate professor emeritus at Medill IMC, and the participating news organizations.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

1 Left censoring is where we do not know when a subscription started, and thus, how long the subscriber has been active. This is a problem for at least two reasons. First, the probability that a customer will churn depends on how long the customer has been active, which is normally accounted for with the baseline hazard function in survival analysis. Second, less loyal customers will churn earlier and be purged from databases, leaving a disproportionate fraction of more loyal customers. Churn rates appear to drop with tenure, but it could be this sorting effect. By selecting only subscribers who started after some date when we have complete data, we avoid these problems.