Professor Mingdi Xin's study explores how a customer’s uncertainty about the value of a software application prior to adoption influences a vendor’s optimal pricing strategy and may explain why perpetual licenses are more profitable than durable-goods theory would suggest.

Why Are Software Subscription Models Falling Flat? Customer Value Uncertainty Plays a Key Role

July 27, 2020 • By Alex D. Bennett

Because software can be used repeatedly without diminishing its value, standard economic theory treats it like a durable good. A conventional durable-goods vendor can optimize its profits by prioritizing subscription-based pricing. But software behaves differently than standard theory would suggest. Although durable-goods theory holds that software vendors should operate under a subscription model, in practice subscriptions have not shown clear advantages over the perpetual licensing model, even as its popularity continues to rise.

The contrast between practice and theory in software licensing is the focus of UCI Paul Merage School of Business Assistant Professor of Information Systems Mingdi Xin’s recent work. Her study, “The Impact of Customer Valuation Uncertainty on Software Licensing,” was published in MIS Quarterly. The research explores how a customer’s uncertainty about the value of a software application prior to adoption influences a vendor’s optimal pricing strategy and may explain why perpetual licenses are more profitable than durable-goods theory would suggest.

Software licensing: theory versus practice

Over the last few decades, software licensing has undergone a series of evolutions. Until the mid-2000s, perpetual licensing was the overwhelming standard. The perpetual licensing norm contrasted with economic theory, which consistently suggested that software vendors could achieve better profitability using a subscription model. Despite the theoretical consensus, subscriptions rarely generated significant revenue for vendors.

Technical advances have made subscription models more common. The new Software-as-a-Service (SaaS) model, which bundles a software subscription with other web-based services, has grown rapidly. Given the excitement around SaaS, one might expect it to drive vendor revenue growth. But the real-world experience has not been so clear. Many vendors switched to the SaaS model but are now pushing for multiyear commitments from customers, transforming a nominal subscription into something resembling a perpetual license.

Xin explains, “There’s a lot of confusion in the field, because the historically prevailing practice differs so much from theory. Now the new SaaS model appears ready to disrupt the marketplace, but vendors have shown reservation. What is going on?”

Valuation uncertainty offers a compelling explanation

The study locates a potential source of the problem in the nature of customer demand: Customers usually cannot fully evaluate a software application before adopting it. To reach a firm valuation, a customer would need to use the software for a long time, integrated into the customer’s wider IT systems and workflows. Reviews or product demonstrations can only tell part of the story. The cost of implementation often complicates the customer’s choice.

In a dynamic game theoretic model, the study examines how customers’ value uncertainty may affect a vendor’s optimal licensing strategy. It shows that given value uncertainty, customers and the vendor face a trade-off. From the customer’s perspective, a perpetual license offers the advantage of a one-time cost at the risk of making an irreversible, bad choice. A subscription provides customers with the flexibility to cancel or modify their plan. But subscriptions come with the risk of price increases over time, potentially increasing the software’s overall cost.

While customers grapple with how to value prospective new software, vendors must make a pricing choice between perpetual licenses and subscriptions. With subscriptions, the vendor has the flexibility to change prices over time. It may charge a low initial price so customers can use the software and resolve valuation uncertainty with a low upfront financial commitment. After customers learn their true valuation, some are willing to pay more, while others may only be willing to pay less for a renewal. The vendor can raise the renewal price to gain from the improved willingness-to-pay in the high-end market, although it would lose demand from lower-end customers.

A perpetual license prevents the vendor from increasing prices later on, but with the benefit of greater up-front revenue. Moreover, customer value uncertainty leads to a less diverse demand distribution, which helps make perpetual licensing more profitable. Specifically, even though customers may research software prior to initial adoption, they may incorrectly assess their perceived valuation, while the actual benefit of the software only becomes clear through use. When all customers take this risk into consideration, their willingness-to-pay for the software prior to adoption becomes more uniform. The study shows that such a reduction in demand diversity can make perpetual licensing more profitable.

Toward a more profitable approach

The study examines various models. In models that leave out customer value uncertainty, the study’s prediction is consistent with conventional durable-goods theory. Adding customer value uncertainty into models produces results that more closely resemble real-world strategies. The vendor’s optimal licensing strategy depends on the nature of the software. Vendors of software that faces high customer value uncertainty and high implementation costs such as complex enterprise software should prefer perpetual licensing. But when subscription is optimal, the study recommends a low-then-high variable pricing path, with the low initial price aiming to encourage adoption and help customers resolve value uncertainty. This finding differs from conventional economic theories, but is consistent with many software vendors’ practice. Another alternative, which may be more profitable for software with low implementation costs and wide differences between high- and low-end customers, is to offer subscriptions at different rates depending on the length of a customer’s commitment.

The study also examines the social welfare impact of customer valuation uncertainty. On one hand, the customer’s initial value uncertainty may produce a net social benefit by reducing the dispersion of demand distribution in the marketplace and leading to more customers adopting the software. On the other hand, due to value uncertainty, a customer may make a bad choice based on incorrect valuation assumptions, resulting in a net decrease to welfare. The overall social welfare effect depends on the trade-off between these two opposing forces. In some cases, a vendor may have incentives to address customer valuation uncertainty if it hurts social welfare outcomes.

Xin hopes her study will help software vendors better understand how features of customer demand may affect the trade-off between different licensing strategies. Her explanation for why software licensing deviates from standard durable-goods theory offers vendors three important insights. First, it explains why perpetual licenses can be optimal in some situations, and why they have historically dominated the software marketplace. Second, the model finds that when subscription is optimal, customer value uncertainty makes low-then-high pricing a more profitable approach than a fixed-price subscription. Third, sometimes a menu of subscription options varying in subscription duration and prices may be optimal.

Mingdi Xin is an assistant professor of information systems at The Paul Merage School of Business. An applied economist with a PhD from the New York University’s Stern School of Business, she conducts research in product and pricing strategies for digital goods and services, IT investment strategies and their competitive implications, and online consumer product search behavior. Her research has been published in Management Science, Information Systems Research, MIS Quarterly and Harvard Business Review.