User Acquisition and Engagement in Digital News Media

Generating revenue has been a major issue for the news industry and journalism over the past decade. In fact, vast availability of free online news sources causes online news media agencies to face user acquisition and engagement as pressing issues more than before. Although digital news media agencies are seeking sustainable relationships with their users, their current business models do not satisfy this demand. As a matter of fact, they need to understand and predict how much an article can engage a reader as a crucial step in attracting readers, and then maximize the engagement using some strategies. Moreover, news media companies need effective algorithmic tools to identify users who are prone to subscription. Last but not least, online news agencies need to make smarter decisions in the way that they deliver articles to users to maximize the potential benefits.

In this dissertation, we take the first steps towards achieving these goals and investigate these challenges from data mining /machine learning perspectives. First, we investigate the problem of understanding and predicting article engagement in terms of dwell time as one of the most important factors in digital news media. In particular, we design data exploratory models studying the textual elements (e.g., events, emotions) involved in article stories, and find their relationships with the engagement patterns. In the prediction task, we design a framework to predict the article dwell time based on a deep neural network architecture which exploits the interactions among important elements (i.e., augmented features) in the article content as well as the neural representation of the content to achieve the better performance.

In the second part of the dissertation, we address the problem of identifying valuable visitors who are likely to subscribe in the future. We suggest that the decision for subscription is not a sudden, instantaneous action, but it is the informed decision based on positive experience with the newspaper. As such, we propose effective engagement measures and show that they are effective in building the predictive model for subscription. We design a model that predicts not only the potential subscribers but also the time that a user would subscribe.

In the last part of this thesis, we consider the paywall problem in online newspapers. The traditional paywall method offers a non-subscribed reader a fixed number of free articles in a period of time (e.g., a month), and then directs the user to the subscription page for further reading. We argue that there is no direct relationship between the number of paywalls presented to readers and the number of subscriptions, and that this artificial barrier, if not used well, may disengage potential subscribers and thus may not well serve its purpose of increasing revenue. We propose an adaptive paywall mechanism to balance the benefit of showing an article against that of displaying the paywall (i.e., terminating the session). We first define the notion of cost and utility that are used to define an objective function for optimal paywall decision making. Then, we model the problem as a stochastic sequential decision process. Finally, we propose an efficient policy function for paywall decision making.

All the proposed models are evaluated on real datasets from The Globe and Mail which is a major newspaper in Canada. However, the proposed techniques are not limited to any particular dataset or strict requirement. Alternatively, they are designed based on the datasets and settings which are available and common to most of newspapers. Therefore, the models are general and can be applied by any online newspaper to improve user engagement and acquisition.