Studying the world’s software development trends can help CIOs make the grade

CIO Magazine, February 1, 1997 By Howard A. Rubin
(http://www.cio.com/archive/020197/metric.html)

How much should you spend to outfit a software development team in India? Will you get better service from a programmer in Russia or Japan? Are your Canadian code writers still productive, or would a Chinese company be more useful to you?

Maybe 10 years ago the information technology chief in a U.S. company didn’t have to bother his head with such arcane questions. But today, CIOs everywhere must master global software economics. Why? Because nowadays, software development is proceeding across all time zones of the world 24 hours a day, seven days a week. The flow of development work from the United States and other developed Western nations to India and Russia and the movement of software products over the Internet illustrates software development’s international scope. For CIOs who can understand and exploit that trend, fame and fortune await. For those who can’t, disaster may be imminent.

Indeed, some pioneering IS executives in banking and telecommunications already are using their knowledge of global software and information technology economics to create internal free market software economies for their own companies. They are forming global software alliances, making strategic offshore investments in developing nations and, most importantly, carefully calibrating the cost and performance parameters of their global software development teams. Software development and maintenance work flows freely to whomever can deliver the most efficient and effective product regardless of time zone or national border. It’s enough to make any Adam Smith devotee proud.

From an internal business viewpoint, the emerging need to manage software development teams in all corners of the globe will require companies to gather more refined development information and more accurate software metrics to support wise decision making and effective management. Alas, the traditional software metrics are woefully inadequate for the task of managing global software development projects, although academia and corporate America are working to understand the status of software metrics and to eventually develop new numbers.

For example, in 1995 I conducted a study of global software engineering companies’ performance along with researchers Ed Yourdon and Heidi Albrecht Battaglia. The study was dubbed the Worldwide Benchmark Project and was funded by Industry Canada, the Canadian equivalent of the U.S. Department of Commerce. Our findings clearly demonstrate the variability of software engineering performance at the corporate, regional and national levels in terms of productivity, quality and capability. We found that based on those criteria, the top-performing software-producing companies in the world are 200 times to 600 times more efficient than the bottom-tier producers. Software quality, as measured by the number of defects per line of computer code, varies by as much as 100 times to 1,000 times among the top and bottom producers.

We also discovered that though those huge disparities in software efficiency are growing rapidly, organizational knowledge about the software function remains sparse. For example, less than one in five companies studied have a complete application inventory. Less than one in 30 know the rate at which their software portfolios are changing. If those numbers aren’t enough to keep you up at night, consider this terrifying finding: Less than one in 100 companies collect consistent quality information about their software projects.

Of course, the world’s nations haven’t done a very good job of keeping track of changing economic conditions either. But if one considers software engineering a core competence of the information age, then calculating statistics about national computing power and telecommunications lines per capita, for example, will be necessary in order for countries to maintain or advance their competitive position. Yet industrial-age measures of performance such as gross domestic product and industrial production continue to be the primary gauges of Wall Street and White House economists. Better predictors of a nation’s growth and potential can be found by looking at a country’s information processing capacity and capability. The country that measures those first will have a leg up in the global competition for jobs and wealth.

Likewise, for corporate CIOs, the race is on. In order to remain competitive, today’s corporation must develop an understanding of the dynamics of global software economics. Doing so will require constant competitive calibration. That, in turn, will point to new forms and uses of benchmarking. Data and information will be collected in real-time.

New metrics, new methods and new competencies are clearly needed in the rapidly changing and increasingly global field of software development. Those CIOs who fail to understand international software metrics will fall victim to dynamics they do not understand. Howard A. Rubin chairs the computer science department at New York’s Hunter College and is CEO of Rubin Systems Inc. He can be reached at 71031.377@compuserve.com.

Pop Quiz

How ready are you to do battle in the global software development arena? Test your knowledge of global software economics.

  1. India and China employ the most software engineers in the world. True or false?
  2. Which country has the world’s most productive software engineers, the United States, Japan, Canada or India?
  3. Which country charges the lowest rate for a three-minute international phone call?
  4. By the end of the decade, will the United States lose or gain in share of worldwide computer processing power?
  5. What country do most global executives think does the best job of supporting business with its technological infrastructure?
  6. Say a market basket of computer equipment and software (PC, printer, notebook, online subscription service and software suite) costs $13,000 in the United States. In India, would the same setup cost $5,000, $13,000 or $20,000?
  7. Switzerland has the most telecommunications lines per capita. True or false?
  8. If U.S. companies outsourced all their software development to low-wage countries like India and Russia, what percentage of the U.S. work could all those countries together handle?
  9. China enjoys the world’s fastest growing economy. True or false?
  10. Which country has the world’s least expensive software professionals?

Answers:

  1. False. The correct answer is the United States and Japan.
  2. Canada
  3. Brazil
  4. Lose. The United States’ share is forecast to fall from its current level of 37 percent of worldwide MIPS to 30 percent by 2000.
  5. New Zealand
  6. $20,000
  7. True
  8. 30 percent
  9. True
  10. Russia

Sources: 1995 Worldwide Benchmark Project and the IMD 1996 World Competitiveness Yearbook. Compiled by Howard Rubin.