Editor’s note: This blog post is the third in a series on key performance indicators (KPIs) that are helpful for media companies and other subscription or membership-supported organisations. The first post was about total customer lifetime value (TCLV). The second post was about core relationships.

Mather Economics has long-term relationships with many of our news media clients going back 10 years or more. We have observed the dramatic change in revenue streams as advertising migrated from print to digital platforms and digital news products were created and sold to subscribers. Through the prism of this experience, we developed several KPIs specific to renewal price changes measuring the effectiveness of a pricing strategy and campaign.

These renewal pricing KPIs include incremental stops, net incremental revenue, reverts above and reverts below, and upgrades/downgrades. Another important metric we developed is the ratio of the net and gross rate changes. In other words, the realised price increase on those customers that remains active following the pricing action divided by the gross price increase applied to those customers.

As an example, if a publication gave its subscribers a US$1 per month increase and the remaining customers’ average rate increased US$0.80, the ratio would be $0.80/$1, or 80%. We call this metric yield.

Mather offers a dynamic subscriber renewal pricing service called market-based pricing. The goal of this service is to improve the economic efficiency of renewal pricing by: a) increasing the revenue generated from subscribers with high demand for the product, and b) decreasing volume loss by avoiding aggressively pricing subscribers with higher sensitivity to price.

The merits of market-based pricing in the publishing industry are well documented, with hundreds of publishers across the world having successfully implemented dynamic pricing strategies over almost two decades.

It is typical for this type of pricing strategy to decrease price-related subscriber loss by 50%, which increases the yield of a renewal price increase substantially. In this era of increasing reliance on subscriber revenue, improving the performance of renewal pricing in generating revenue and minimising subscriber churn is vital to the success of publishers.

To support dynamic renewal pricing, Mather develops econometric models to estimate the retention propensity and price sensitivity of individual subscribers. We then recommend renewal rate adjustments according to the relative distribution of price sensitivity across subscribers in the market, and we validate the performance of the predictive models through A/B testing, where a statistically representative control group of subscribers is excluded from the price increase. Comparing the performance of the test and control groups enables us to calculate the pricing KPIs described earlier.

We have found tenure is an important predictor of price elasticity. The first year of a subscriber’s relationship with the publication is when he is most likely to stop due to a price increase. If he reaches his third year, his renewal price elasticity will drop significantly. So, for publishers, when it comes to renewal pricing power, time is on your side.

Declining yield

Yield has become an increasingly important KPI for our pricing programmes. One key finding of our research, especially for those markets that have implemented several rounds of renewal pricing (both dynamic renewal pricing and standard, across-the-board pricing), is that the yield on renewal pricing generally declines with each renewal price increase. This is an intuitive result for many who have worked in audience and circulation groups, and this customer response should be anticipated when developing a pricing strategy.

To demonstrate this finding, the graph below from one of our partners illustrates the average yield on pricing has gone from 79.2% during the first renewal pricing instance to 65.4% several years later during the sixth pricing instance. Importantly, the average stop rate over this same period declined significantly.

The implication being is that while fewer subscribers stopped their subscriptions during a renewal pricing instance, fewer subscribers were accepting the full amount of the rate adjustment.

Yield analytics

To understand the factors affecting yield, our team ran hundreds of model specifications on a comprehensive data set including customer transactions and detailed performance results from prior pricing initiatives. We also included macroeconomic indicators to determine if the state of the economy influenced renewal pricing response.

Some of those inputs are listed in the figure below, along with their directional impact on yield. For example, we found the yield on a subscriber’s previous pricing instance was negatively associated with the yield on his next instance, suggesting subscribers who do not accept a price increase (but retain their subscription) will tend to similarly reject future pricing attempts.

We also found the presence of previous price-related customer service interaction (reverts) was associated with lower future yield on pricing. Finally, we found macroeconomic indicators to be significant predictors of yield, with higher unemployment and price levels in the economy being correlated with lower future pricing yield, while higher personal savings rates tended to improve future yield.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .