This case study explores the trials and successes experienced by a small, 20 employee, internet start-up, Kagi.com, in its transition to Unicode to enable global electronic commerce. It then explores suggested strategies devised based on that experience.

Kagi is a 100% internet based payment processing service that started 4 1/2 years ago using legacy ASCII based software. They quickly grew to the point where they now service over 4,500 vendors from six continents with customers from seven continents. At this point, they have completed over 1.7 million transactions in multiple currencies, with over half their sales coming from Europe and Asia.

To accommodate the rising level of global sales, Kagi made a decision in 1997 to standardize their systems on Unicode. Using commercial off the shelf software, servers, browsers, and Java, they created a Unicode based payment processing system. But there have been many trials in establishing the system, particularly with respect to the learning curve, character set conversion errors, tools incompatibility, legacy data, and generic issues around e-mail responses with vendors and customers with unknown mailers.

This paper will present each of the above problems in how they were encountered and how, or if, they were solved. It will then present a high-level view of the architecture as it evolved with suggestions for future strategies.