Think Big: Introducing Product Line Engineering in a Small Company

Author: Dr. Martin Verlage, MARKET MAKER Software AG
February 2004

In the year 1999, MARKET MAKER, a small Kaiserslautern Germany based company that develops software for stock market data information systems and provides services to retrieve the data for applications, faced the emerging need of stepping into the market of Internet services. Given the need to maintain relationships to a broad spectrum of existing customers, the question was which market to address first. The answer was simple: all at once!

The next step was to pick a method that addressed the needs of a small company when creating a family of products. Maybe MARKET MAKER wouldn't have introduced a systematic way to deal with multiple products if we were aware of the common thinking about product line engineering (PLE), which is that PLE is a big process and suitable only for big organizations. We looked at it from a different angle: the PLE motivation fit (e.g., lower cost, less maintenance, faster time-to-market), the staff is well educated (e.g., many “software engineers”, few “programmers”), and the business goals were met by the promises of PLE.

A project was conducted which resulted in the “i* ProductLine”. This product line is a set of components making up applications to process financial data from stock markets and qualitative information such as news and company profiles. The applications usually run as web services or web sites to deliver data to customers who pay for the data access. The market segments vary from solutions for bank clerks, XML document servers, to publicly available web sites for online brokering companies.

Today, dozens of web services run at customer and MARKET MAKER sites to deliver HTML pages and responses to requests from many clients. PLE helped to create a business model that is supportive and open to future markets. Creating new products relies on a repository of domain-specific assets, helping to reduce development time by more than 50% and reduce cost by roughly 70%.

1 2 . . Next page >