Customer perceptions of a thin-client micro-payment system: issues and experiences.

ABSTRACT

Two fundamental payment methods exist far on-line information purchase: macro-payment and micro-payment. Traditional macro-payment methods, like credit and charge cards and digital currency, are suitable for large-value, low-volume transactions. However, large-volume, low-value commodities, such as discrete units of information from a Web site, better suit a micro-payment model. In micro-payment, customers pay for large numbers of small value goods (e.g. per-Web page view) with "e-coins ", typically of very small value each. We have carried out an empirical assessment of micro-payment and macro-payment purchasing models for an on-line newspaper application. We report on the design of our experiment, the two kinds of micro-payment (client and server-side e-wallets) used, and customer feedback. We also carried out an assessment of customer effort and economic trade-off when using these services and compare the results of this assessment to a survey of customers using each system. We present directions for further on-line payment research aiming to improve the overall satisfaction and efficiency of payment models far end-users.

Keywords: micro-payment; usability; e-commerce; e-money

INTRODUCTION

Most current e-tailing systems adopt a macro-payment model and architecture. A user makes a small number of on-line purchases that have a reasonably high purchase price. In order to pay for these purchases, a "heavy weight" interaction between the vendor of the product or service and an authorisation agent (bank, credit-card company, etc.) system is carried out. This typically involves the user supplying credit card details or "digital money" certificates, which are communicated to the authorisation system using complex encryption algorithms. Business processing logic and database updates are performed by the authoriser before the purchase is approved. The vendor system waits for approval before providing the customer with goods or services. This approach works well for relatively small numbers of transactions and relatively high purchase price (to offset the cost of authorisation) (Dai et al., 2001). However, in some e-commerce scenarios this approach has a number of fundamental flaws. It requires the authorisation system to always be on-line. High numbers of transactions or low-price purchase items are infeasible, due to bottlenecking or prohibitive cost per-transaction. In addition, with most approaches the customer's identity cannot generally be hidden from the vendor. For example when using a subscription-based approach i.e. a single macro-payment for on-going supply of services, the vendor must be supplied with customer-identifying information. In addition, if the customer only makes use of a small fraction of subscribed services, they spend a comparatively high amount of money for what they use. Most subscriptions are to a single vendor's service and don't cover the purchase of low-value commodities from multiple vendors.

We describe the NetPay micro-payment model and architecture we have been developing. NetPay provides an off-line micro-payment model using light-weight hashing-based encryption. A customer buys a collection of "e-coins" using a macro-payment from a broker. These coins are cached in an "e-wallet" (stored either on the customer's machine or on broker and vendor server machines). The customer, when buying many small-cost items from a vendor, pays for these transparently by the passing of e-coins to the vendor. Periodically the vendor redeems the e-coins with the broker for "real" money. E-coins can be transparently exchanged between vendors when the customer moves to another site.

In this paper, we give an overview of the concept of micro-payment versus macro-payment models of e-payment. We outline our research methodology of assessing the perceptions of customers of micro-payment versus macro-payment models for e-tailing systems. We present he software architecture and design for NetPay, a new micro-payment system we have prototyped, for deployment with thin-client vendor interfaces for customers. We describe three kinds of experiments we have done on our NetPay prototype, to assess micro-payment versus macro-payment usability, performance and overall qualitative characteristics for e-tailing systems payment. We compared two kinds of NetPay-based micro-payment systems (client-side wallet and server-side wallet) and a subscription-based macro-payment system. We conclude with an outline of our further plans for research and development in this area.

MICROPAYMENT FOR E-TAILING SYSTEMS

Consider the scenario of customers wanting to browse on-line newspapers (Dai et al., 2001). Using the typical approach of subscription-based payment, the user would first have to subscribe to the newspaper by supplying personal details and payment details (credit card number, etc.). The newspaper system would then make an electronic debit to pay for their subscription by communicating with an authorisation server. The user would then normally go to the newspaper's site where they login with an assigned user name and password. The newspaper looks up their details and provides them access to the current edition if their subscription is still current. If the user's subscription has run out, they must renew this by authorising a further macro payment from their credit card. Figure 1(a) outlines the key interaction use cases for this scenario. Problems with this approach are that there is no anonymity for the user (the newspaper system knows exactly who they are and when and what they read), they cannot browse other newspapers without first subscribing to them too, and they must pay for the whole newspaper, even if they want just one or two sections or articles These issues apply to many other information sources on the Internet where vendors want to charge for content (Blankenhorn, 2001; Herzberg, 1998).

[FIGURE 1 OMITTED]

An alternative approach is to use a "micro-payment" model. There are several approaches to micro-payment (Furche and Wrightson, 1996; Herzberg and Yochai, 1996; Hwang et al, 2001; Manasse, 1995; Rivest and Shamir, 1997; Stern and Vaudenay, 1997). We outline the basic interactions of the NetPay model we have developed (Dai and Grundy, 2002). Figure 1(b) outlines the key interaction use cases for this scenario. The user first goes to a broker and purchases "e-coins" using a single macro-payment. These are stored in an "e-wallet", either on the user's machine or on the broker server. The user can then visit any vendor site they wish, for example an on-line newspaper. Each time they need to purchase a small-value item e.g., view an article (or section or …