Publishing Executive Blog
4 November 2019
By Matt Lindsay
Companies with subscribers can apply different pricing across customers to a degree not possible in many industries because they have one-to-one relationships with each customer. This one-to-one communication enables personalization and customization of the relationship between customer and company across many dimensions, pricing included.
Other industries, such as consumer packaged goods, have prices that are observed by everyone in the market, and publishers also face this issue when they are acquiring new customers. Having multiple acquisition prices in the market may cause confusion or channel conflict, but renewal pricing can be done individually with each customer because the renewal price communication is not public. Economists call this first-degree price discrimination.
But just because something can be done does not mean it should be. In this article, we will argue that the benefits of pricing variation across customers far outweigh the costs of this strategy for publishers AND their customers.
Knowledge is Power
We know from hundreds of price tests in the US, Canada, Australia, New Zealand, Scandinavia, Northern Europe, and Latin America that certain customers are 20 times more likely to stop their subscription after a price increase than others, and we know the customer characteristics that identify someone as price-sensitive. A one-price model is clearly not the best strategy for publishers, who can, much like other industries such as airlines and hotels, use sophisticated pricing analytics to increase revenue and operating margins.
Mather Economics develops econometric models of customer retention and price elasticity using data from publishers on their customers’ actions following prior price changes. These models are effective at measuring the influence of individual factors on a customer’s decision to renew their subscription or stop the service. In addition to renewal price increases, other factors are important in predicting whether a customer will pay a renewal notice, such as their tenure as a subscriber, their acquisition channel, their original offer term, their payment method, their approximate income level, and their approximate age group.
Once an econometric model has been developed, a test of the model’s predictions on price sensitivity is implemented by selecting a statistically valid sample of subscribers and changing their renewal prices while keeping a comparison group of subscribers at the same price. This technique is called A/B testing: the customers that receive the price change are the target group, while the customers that are kept on their same price are the control group. Observing these groups over time, we track how many customers pay the new price, how many do not pay their bill, how many customers call customer service, and the number that change their service level.
The difference in customer behavior between these two groups can be attributed to the price change since it is the only factor not held constant. Both groups will have the same mix of customers by service, price, tenure, income, age, geography, and other relevant characteristics. If 5% of the target group stops their subscription and 3% of the control group stops, we assign the additional 2% of stops to the increased renewal price.
The two charts above show the different stop rates between target and control groups across customer tenure groups. The chart on the right shows that customers in the target group that are in their first year of subscribing had a 27.6% stop rate after a price change (labeled BAU for “business as usual”). The control group customers in their first year that did not receive a price increase (labeled “Holdout”) had a 18.3% stop rate. The chart on the left shows the price elasticity, the ratio of the % of stops divided by the % change in price, for each tenure group. Customers in their first year of subscribing are about 10 times as sensitive to price as customers who have subscribed for more than five years.
This knowledge enables publishers to do two things:
- A) avoid charging price-sensitive customers renewal prices that would cause them to stop their subscription, and
- B) increase prices to those customers who value the product so much they are willing and able to pay more for their subscription.
To apply this insight to subscription pricing operations, Mather Economics developed a recurring process where subscriber data is received from newspapers weekly, and suggested renewal prices are provided for each subscription that is receiving a renewal notice. Statistically representative control groups are used to measure the effect of prices on retention and revenue, and we adjust the recommended prices based on the observed incremental stop rates across customer segments.
In some countries there are restrictions on the variation in subscription pricing that can occur across customers. These restrictions can be from legal limitations or regulatory standards. Regulatory agencies have dropped most limits on pricing variation as the news media industry has struggled with print subscription volumes. Legal restrictions on subscription pricing are rarer and do not affect publishers in most countries.
With Power Comes Responsibility
Publishers often feel that variable pricing by customer is deceptive or dishonest, but, gradually, more publishers are observing a variable pricing strategy benefiting their customers AND themselves. Without the flexibility to target price increases, publishers have to raise subscription rates for many of their most price-sensitive, least-affluent, and youngest subscribers at the same level as they do to their wealthiest, least-price-sensitive customers, and these vulnerable customers frequently stop their subscriptions while the other customers would probably have paid more. Variable pricing strategies can reduce price-related customer losses by as much as 75%.
You may wonder how charging different prices for the same product works in a business built on long-term relationships and in a competitive environment where customers can go elsewhere. It really depends on local customs and what is the right choice for your organization and for your customers.
Arguably, investigative journalism is a public good that benefits everyone once it is produced. Differentiated subscription pricing benefits society by providing greater financial support for journalism and greater access to independent news media.
More on Regression Analysis and A/B Testing
As discussed, regression analysis can measure the effect of important factors on subscriber retention in isolation. Once these models have been developed, predictions can be made on the number of accounts that will stop due to a renewal price increase and who is most likely to stop. Using A/B testing with control groups to validate the accuracy of the model predictions is the best way to establish a closed-loop pricing and retention optimization process where insights from price changes are incorporated into future changes, thus improving the efficiency and performance of the pricing strategy.
With digital subscriptions, the number of variables that can be included in a regression model of retention and price elasticity increases substantially due to the availability of data on customer’s engagement with the content online. We can identify who has accessed content, what they read, on what platform, and at what time. As expected, subscribers who are more engaged with the content are less likely to stop their subscription following a renewal price increase.
Several predictive metrics are consistently important in our models of print and digital subscriber retention. A customer’s active life with the subscription, often called their tenure, is often the most significant predictor of price elasticity as illustrated in the A/B test results. The customer’s engagement with the product is another very important predictor of their propensity to churn due to a price increase. We can classify variables that are important for price elasticity into consumption, interaction, attitudinal, time, and socioeconomic categories, and there are other metrics that do not fall into these groups.
- Consumption metrics describe the quantity, frequency, and time spent with the content.
- Interaction metrics describe actions taken by the customer while on the site or while they are engaged with the content.
- Attitudinal metrics are those that measure the level of enthusiasm or loyalty an individual has for a topic or community.
- Time metrics reflect when events occurred during a subscriber’s lifecycle, and the overall time of activity on the account, often called the account tenure.
- Socioeconomic metrics include factors that characterize an individual’s demand for the product, such as disposable income, price sensitivity, age, gender, macroeconomic indicators, education, and household type.
Pricing Strategy Is Key to Success
Changes to the advertising industry make it likely that most revenue for news publishers will come directly from their audiences for the foreseeable future. A one-size-fits-all pricing strategy does not help customers who need lower prices, and it does not help publishers who need revenue from customers who are willing and able to provide it. Publishers who use intelligent subscription pricing models will be those who survive using a subscription-revenue business model.
