Pricing financial services innovations

Abstract

The number of innovative financial solutions introduced to markets has grown considerably in the past decade owing to emerging digital technologies, deregulation and market fragmentation. Examples are abundant in the worldwide markets for insurance, credit products and transaction processing services. A question of growing interest is how firms should price these innovations. The optimal introductory pricing of financial innovations may vary as a function of factors such as price sensitivity of the market and competitors’ ability to introduce competing financial solutions. In this article, we examine the role of these factors in the optimal pricing of a financial innovation. Using an agent-based simulation framework, introductory pricing strategies that maximize profitability under various market conditions are identified. Results indicate that lower levels of market price sensitivity and longer time horizons for competitive entry create pricing opportunities for financial innovators. However, the relationship becomes more complex as market price sensitivity increases or competitive market entry becomes more immediate. Detailed recommendations for optimal pricing of financial innovations under various market conditions are provided, and the article concludes with strategic recommendations for pricing innovative financial services.

APPENDIX

Selection of parameter values for the simulation model

Price, price elasticity and time horizon

The simulation examines diffusion processes resulting from sales prices of 0.4, 0.7, 1 (base level), 1.3 and 2. In addition, price elasticity is systematically varied at four levels: 0.2, 0.6, 1 and 2. These values are in line with those used in previous theoretical studies (for example, Lehmann and Esteban-Bravo, 2006) and those found in meta-analysis of empirical studies (Tellis, 1998; Bijmolt et al, 2005).

To capture the expected time of competitive entry, we applied the customer groups discussed by Rogers (2003) in the base model. Rogers categorizes adopters into five categories: innovators, early adopters, early majority, late majority and laggards. We examined the time horizons of 3, 6, 9 and 12 periods. These numbers represent the average time when approximately 3 per cent, 16 per cent, 50 per cent and 84 per cent of the market have adopted the product in the base case. These values approximately represent the cumulative adoption of the four first groups of the aforementioned categories of adopters as discussed by Rogers (2003). Furthermore, because increasing price may lead to slower adoption rates, and hence longer diffusion patterns, we also captured the dependent variable at a time when 95 per cent of the market have adopted the product. This allowed for examining conditions where the firm does not have any concern with the amount of time it may take for the diffusion process to complete and the goal is to maximize profits regardless of the time horizon.

Parameters p and q

The values of parameters p and q are based on the values estimated by previous empirical research and meta-analyses. As this study seeks to examine the context of financial innovations, the choices of values for parameters p and q represent average values considered for such an innovation in previous research. Thus, we fixed the aggregate-level parameter p to 0.0142 and the aggregate-level parameter q to 0.545, values estimated for the diffusion of online banking (Libai et al, 2009). These values are in line with those used in earlier studies that examined online banking (Hogan et al, 2003). Moreover, previous research has found that consumers perceive greater risk in the adoption of services compared to the adoption of goods, and thus adoption of services relies more heavily on WOM and personal sources of information, and less so on information from marketers (Murray, 1991). Therefore, the choice of the two parameters in this study are in line with the average values estimated in previous meta-analysis studies (Sultan et al, 1990; Jiang et al, 2006).

The aggregate-level values of parameters p and q were converted to the individual-level parameters p _i and q _i using the methods suggested in the literature (Goldenberg et al, 2002; Toubia et al, 2008). The parameter p is the same at both aggregate and individual levels. However, the individual-level value of parameter q is calculated by dividing the aggregate-level value of parameter q by the number of individual ties. Therefore, the overall diffusion process resulting from this model are comparable to those at the aggregate level (Goldenberg et al, 2002).

Fixed variables

We fixed the average number of one-to-one connections between one consumer and other consumers to 14, a value that is in line with the average number found in previous studies (Goldenberg et al, 2007; Libai et al, 2010; Ghoreishi Nejad, 2011). For the purposes of the simulation, we fixed the market size – number of potential consumers – to 3000, a value that is in line with earlier studies (for example, Goldenberg et al, 2007; Ghoreishi Nejad, 2011). Furthermore, because the average number of connections is used in converting the aggregate-level value of q into individual-level values, it is also indirectly captured in the diffusion process through the influence of consumers on each other through q _i , the number of social ties does not significantly alter the results (Ghoreishi Nejad, 2011). It is important to note that we further examined values of 4 and 24 and the results remained the same.