Propensity to subscription

Dynamic paywalls are a growing strategy amongst publishers, providing a tool for balancing between 2 main business KPIs: maximizing paying subscriptions while minimizing the pageviews lost.

Marfeel offers a method in its SDK to obtain the probability that a reader will subscribe if the current page view is dynamically closed.


window.marfeel.cmd.push(['compass', (compass) => {
    compass.predict().then(result => {

          result will be an object like
          { probability: 0.01234543344 }
          from here on, use it as needed.

          myVar = result;

          myPrediction = result.probability;

Threshold decision

The method returns a probability, it’s up to the publisher and the developer to decide the threshold from which it makes sense to start closing pages. This will greatly depend on publisher’s strategy, but as a guidance our research suggests that a threshold of 0.5% (0.005) could yield a Pareto result of retaining 80% of subscriptions while losing only 20% of the traffic, compared to other simpler dynamic paywall strategies.

Depending on publisher’s priorities, it might make sense to take on a strategy with different thresholds depending on section, type or age of content, user profile, or any other variable.

Performance and Tracking

While a generic probability will always be returned, in order to get optimal performance make sure you set a proper instrumentation of the pages closed with paywall, and subscriptions. This way the prediction algorithm will learn how users behave in your site specifically.

Returned probability will be automatically logged as a pageVar and is accessible in raw data, offering then KPIs for the performance of this dynamic paywall, such as:

  • % of users that subscribe when finding a dynamic paywall
  • Comparison of actual percentage vs predicted probability
  • Average predicted probabilities by dimension