Figure 1: The effect of pricing on retention for two different subscribers

The role of data in subscription pricing

By Matthew Lulay, Mather Economics

As our ability to collect, analyze and act on data grows, so does the role it plays in our daily lives. In no area is this more evident than in consumer economics, where data analysis fuels everything from the types of advertisements we see online to the prices we pay at the grocery store.

The important role of data carries over to the publishing industry as well, where data analysis provides decision makers with valuable information across all phases of business operations. In digital advertising, website traffic data combined with advertising sales data allows publishers to understand which sections of the website are more lucrative from either an advertising or subscription perspective, giving them the ability to set the meter level intelligently by section to maximize revenue.

With content, digital traffic data allows publishers to make article recommendations to users on content that is of interest to them in real time, driving engagement with the market and allowing for additional advertising delivery.

Finally, with audience, transaction and demographic data allow publishers to better understand the pricing sensitivity of individual households, allowing for more efficiency in subscription pricing.

In this article, we will explore the growing use of data in publishing and how it is used in the dynamic subscription pricing process.

Traditionally, the world of newspaper subscription price increases revolved around a flat increase, applied to everyone equally, at the same point in time. The increases, generally done to offset declining advertising revenue, were effective in the short run in increasing home delivery revenue, but often were met with swift declines in volume from price-sensitive subscribers. In need of revenue, but faced with unacceptable levels of price-driven circulation decline, newspapers gradually moved to a dynamic pricing model in which product pricing was not the same for every subscriber, but was rather differentiated based on the estimated willingness to pay of the subscriber.

Dynamic pricing (referred to as discriminatory pricing by economists), is not a new concept. Defined as the practice of charging different prices to different consumers for identical goods, dynamic pricing is widely applied throughout modern economies. In fact, if you have ever used a coupon to purchase a product, presented a loyalty card when visiting a favorite business, taken advantage of a senior discount or paid more for a flight because you booked last minute, you have been involved in the dynamic pricing process.

Dynamic pricing is beneficial to publishers, as it allows them to capture additional revenue (consumer surplus) from inelastic (less price-sensitive) subscribers, where the demand from those subscribers is higher for the product, meaning they are willing and able to pay more. Dynamic pricing is also beneficial to consumers, since publishers can charge elastic (more price-sensitive) subscribers a smaller amount for the product, thus preventing them from being priced out of the market by traditional pricing.

This process allows producers to extract more revenue from subscribers who have a higher demand for the product while simultaneously supporting circulation volume by charging less to subscribers with a lower demand for the product. The net effect of dynamic pricing is higher revenue and volume vs. the single-market price method.

We can see how dynamic pricing is beneficial, but for the process to work effectively, publishers need a method for segmenting the market to identify subscribers with different demand for the product. That is where the data-analysis piece comes in, and here at Mather, our market-based pricing program helps publishers apply the dynamic pricing-process to achieve the gains mentioned earlier.

To identify the pricing sensitivity of individual subscribers, Mather leverages four primary data sources:

  • Historical transaction data, which includes several years of circulation starts, stops, restarts, change of service, complaints and payments.
  • Subscription information, which includes data on subscriber rates, frequency of delivery, payment method, tenure and term length.
  • Demographic data, including age and income information.
  • Digital engagement data, which is a significant predictor of retention as well as pricing sensitivity.

Survey, psychographic and macroeconomic data can also be included depending on availability. These data sources are then combined to create a profile of subscriber activity over a several-year time period.

Once the data is assembled, an econometric approach called “survival analysis” is used to estimate the retention probabilities and price elasticities of various subscriber groups.

Survival analysis was originally developed for application in the bio-sciences and is a time-series methodology for predicting the time interval to a particular event. In its application to publishing, we use survival analysis to measure retention at various points in time, and we also use it to measure the shift that occurs in retention for various groups when a price increase is applied. This shift in the curve provides insight into the pricing sensitivity of a group of subscribers.

Figure 1 illustrates the effect, in which two subscribers with the same base retention probability are affected differently by a price increase. While both subscribers retain at 56 percent under no price increase after 12 months, the increase shifts subscriber B’s retention down 15 percentage points while subscriber A’s retention shifts down only 8 percentage points (indicating less sensitivity to price).

We estimate these curves for a variety of different subscriber cohorts to develop profiles of retention and elasticity for every subscriber in a market’s database. Based on these findings, we can then distribute rate increases efficiently, with those with the lowest elasticities receiving more aggressive increases while those with the highest elasticities receive more conservative adjustments. By doing this, we can achieve revenue gains with a smaller impact to volume when compared to traditional flat increases.

As highlighted by the dynamic pricing example here, it is clear that data is continuing to play a growing role in our daily lives and is becoming especially important to provide decision makers with the information necessary to make quality decisions in an increasingly complex business environment. Publishers can take steps to harness the power of data by investing in data collection and analysis tools, hiring and training the talent needed to analyze the data and partnering with high quality vendors when outside expertise is necessary.

Matthew Lulay is director of Consulting Services at Mather Economics, a CNPA Allied Partner. Learn more at mathereconomics.com.

Got an idea for a Thursday Tipsheet? Contact CNPA’s Joe Wirt at (916) 288-6021.