Survival Analysis: A New Way to Predict Subscription Order Retention

Daily newspapers write thousands of new orders every week. These new subscriptions or starts come from a variety of sources like telephone, door crew or carrier solicitations, and from voluntarily motivated orders and resolicitations of previous cancellations (stops). Starts can also be classified by the type of order. Sometimes they are motivated by special subscription inducements or premiums like two months for the price of one (24-1) or money back guarantees. Sometimes no special premium is used at all — a straight order. All of these sources and types of starts have an underlying commonality: their lifespan or longevity is variable. Predicting and controlling this variability is the key to circulation growth. Questions like the following must be addressed: What is the most cost effective order? How many months will it last? Is there a critical point in time when a particular order is most vulnerable to cancellation? How does the source and type of start interact? Can we develop retention baselinedata that provide reasonably accurate circulation forecasts? Clearly, there is a pressing need for answers to these questions. Unfortunately the answers have been slow in coming. Along these lines, Jackson and Stamm lament: