Throughout recent history, paywalls have had one thing in common: paywall access was determined by content rules rather than customer data. Essentially, it didn't matter whether prospective subscribers visited a publisher multiple times, or were first-time visitors-they received the same experience. These were one-size-fits-all models. Our vision was to create something dynamic and personal-something that could make real-time business decisions on access, and recognize where people are in the purchasing funnel in order to serve the right message to the right user. We determined it had to be a customer-led model. We use machine learning and a proprietary algorithm to predict the likelihood of someone subscribing, recognizing where they are in the purchasing funnel. This likelihood to subscribe dictates the paywall experience a customer receives. The process takes place in real-time-up to 15 million times per week. In short, we have built a Dynamic, Propensity-led Paywall.