Read Full Paper

‘Freemium’, a pricing strategy whereby a firm offers a basic product or service at no cost to the user, has attracted considerable attention in recent years. Having led to the rapid rise of companies such as Skype and Dropbox, countless startups have also adopted freemium as their business model. After all, there are obvious reasons why freemium is attractive, such as saving users the hassle of payment, the illusion of advertising as a ‘last resort’ revenue source, and the power of ‘free’ as a behavioural marketing tool. That said, it is a complex strategy, and despite its popularity, successfully implementing it remains a challenge. As one investment manager put it, “Too many freemium models have too much free and not enough mium”.

From the outside, freemium looks like classic product-line screening, in that a company offers a selection of products at different prices to segment the market. However, studies have shown that a profit-maximising firm should always ensure the low-end product’s price is positive and maximise the single-product profit – seemingly the exact opposite to freemium. In light of this, a growing body of literature in marketing and information systems is looking at the more nuanced reasons behind the rise of freemium. Specifically, a recent study by Zijun Shi, Kaifu Zhang, and Kannan Srinivasan investigates market dominant firms’ product and pricing strategy under network effects, with a focus on the optimality of the freemium strategy. They set out to answer two key questions: what are the necessary conditions for the optimality of freemium, and what are the principles that should guide the design of freemium?

To go about answering these questions, the authors build a single-period monopolistic screening framework and study whether and when perpetual freemium remains an effective strategy once a product has achieved sufficient recognition and diffusion-related factors have declined in importance – something which is especially relevant to firms who have almost reached market saturation. More precisely, they ask whether ‘network effects’ from product usage alone can justify the freemium model, when a company’s sole objective is its single-period product line profit. Network effects being that “Consumers’ valuation of a product varies depending on how many other consumers are using the product or compatible products”. These network effects can be created by both direct interactions (e.g., a free user sharing a file with a paid user) and indirect behavioural factors (e.g., a consumer values the paid product more if there are more free product users because the consumer derives social prestige from using the high-end product).

What they discovered was that in a standard screening framework without network effects, freemium is never optimal, and the company always selects the efficient price point for its low-end product, thereby supporting existing literature. The authors prove this using a generalised quasi-linear utility function and type distribution. The study goes on to show that even with network effects, freemium is not usually the optimal model. Specifically, when network effects are identical across products, the company has a greater incentive to expand its network size and might find it profitable to sell to low-end customers. That said, the authors are quick to explain that “This does not mean freemium is an equilibrium strategy”. Rather, the company should offer a low-end product to attract customers, while keeping the its price positive. The study reveals that freemium could indeed be viable if there is sufficient difference in network effects of the high- and low-end products. Put another way, the firm’s product line has to be such that paid users gain access to larger network effects compared to non-paying users. As the authors themselves point out, “Paying for an upgrade should, in fact, be paying for network effects”; it is not simply about the number of features the firm provides. Finally, and in surprising contrast to earlier studies on efficient price and inefficient quality, when adopting a freemium strategy, the findings show that a company should provide a low-end product whose quality is above the efficient level. While this seemingly reduces differentiation, it helps to retain the low-end consumers.

While the study focuses on product-line profit being the main driver of firm strategy, the authors admit that it leaves out a number of behaviour factors that are nevertheless relevant to freemium strategy. As such. they highlight that it would be of interest for future research to extend the current model to include advertising income, competition, and user subsidisation (where negative price is possible) to name a few.