Loss of Productivity During Java Adoption

In addition to the two-month training period, it will take another six to 10 months for trainees to gain experience and mature as Java developers. During this period, COBOL migrants will not operate at the same level as Java "veterans" - those developers whom enterprises might hire or obtain from external service providers (ESPs). Enterprises should take into account this a lack of productivity during the time required for the developer's skills to mature.

"Professional" and "Casual" Java Developers: The maturity period will be different for "professional" and "casual" Java developer categories. Not all COBOL developers trained in Java will become professional Java developers; some will become casual developers. Professional developers use an architectural approach with rigorous analysis and design, and are capable of developing three-tier logic architecture, JavaBeans and Enterprise JavaBeans, and robust enterprise- or global-class quality applications. Casual developers build simple applets and servlets, usually using visual tools. Their area of expertise is usually limited to graphical user interface development and accessing relational database management systems (RDBMSs) via Java Database Connectivity. Casual developers require the same initial training as the professional developers, but their maturity period will be shorter and their loss of productivity will be smaller.

Comparing Developers' Productivity: Almost always, an enterprise has the option of hiring (or renting from an ESP) a developer who is proficient in Java and productive from day one, rather than training a COBOL veteran and waiting until he or she becomes equally productive. To estimate the loss of productivity during the Java adoption period, we compare a COBOL "migrant" with a Java developer who received training and then acquired one year of Java project experience. Figure 4 shows a migrant's productivity measured as a percentage of a Java developer's productivity, i.e., a migrant's productivity is measured in the percentage of time that a Java developer requires to accomplish the same project (not a screen or a piece of code, but an entire project). For example: If a COBOL migrant needs two months to accomplish a certain project, then a hired or rented professional Java developer will need only 20 percent of the same two months' time to accomplish the same project.

Let us assume that a COBOL developer's annual salary is $65,000, with a total annual compensation of $78,000. (This is the average salary for a COBOL senior programmer in the United States; see Figure 3.) Let us further assume that an enterprise can hire a professional Java developer for $78,000 a year and a casual developer for $72,000 (these are average Java salaries nationwide). The total compensation (salary + 20 percent of benefits) will be $94,000 and $86,000 annually, respectively.

We split the Java adoption period into two-month periods. To estimate the loss of productivity, we compare the cost of using a COBOL migrant vs. hiring a Java developer. The Java developer will cost more, but will take less time to accomplish the same project.

The COBOL developer's compensation per month: ($78,000/12) = $6,500. The hired professional Java developer's compensation per month: ($94,000/12) = $7,800.

Example of a Loss Calculation: A COBOL migrant needs two months to accomplish a particular project. An enterprise will spend on the developer's compensation: $6,500 per month x two months = $13,000.

If an enterprise instead hires a professional Java developer, the developer will need only 20 percent of the same two-month period to accomplish the same project. An enterprise will spend on the developer's compensation: $7,800 per month x (two months x 20%) = $3,100.

Hiring a Java developer will save an enterprise: $13,000 - $3,100 = $9,900 on the duration of two months. By the same token, an enterprise will lose $9,900 if it uses a COBOL migrant instead of a hired Java developer.

An enterprise will lose $20,000 if it uses a COBOL migrant instead of a hired Java developer on the duration of the 10-month period that a COBOL developer needs to become a fully productive professional Java developer.

A COBOL developer's compensation per month: ($78,000/12)* = $6,500. A hired professional Java developer's compensation per month: ($86,000/12) = $7,200.

An enterprise will lose $10,000 if it uses a COBOL migrant instead of a hired Java developer for the duration of the 10-month period that a COBOL developer needs to become a fully productive casual Java developer.

As Figure 5 and Figure 6 show, using a COBOL migrant instead of a professional Java developer will result in $20,000 in losses. Using a COBOL migrant instead of a casual Java developer will result in $10,000 in losses.

Retaining a Trained Developer

After providing training and allowing for their developers' maturity, enterprises face a serious risk of losing their developers to other opportunities. The job market demand for Java developers is so high, and salaries and benefits paid by ESPs and dot-coms are so lucrative that COBOL migrants will be tempted to migrate again - this time to new employers. To retain these employees, enterprises should be creative with social and monetary incentives. A minimum salary raise of not less than 20 percent for professional developers and 10 percent for casual developers at the end of the training/maturing period is a prudent precaution, as this increased salary would rarely match the salary these developers could receive if they leave (see Figure 7).

Total Cost of Migration

For the total expense of training and retaining a COBOL developer, see Figure 8. For a comparison between previous and current compensation, see Figure 9.

As Figure 8 shows, transforming a COBOL developer into a professional Java developer will cost an enterprise approximately $57,000 in expenses and losses, an amount equal to almost 90 percent of a COBOL developer's salary. Transforming a COBOL developer into a casual Java developer will cost an enterprise approximately $39,000 in expenses and losses, an amount equal to almost 60 percent of a COBOL developer's salary.

Note on Amounts Used in These Calculations: These calculations are based on the assumption that for a senior programmer, the average COBOL salary is $65,000 and the average Java salary is $78,000. This COBOL and Java salary information (based on the title of senior programmer) was obtained from the following sources:

• Realrates.com, which dynamically captures salaries and consultant hourly rates by language or specialty or role or region (see www.realrates.com)
• The 1999 National Data Information Technology Salary Survey (see http://www.romac.com/)

Gartner's Application Development Commentary COM-11-8519, 20 September 2000.

Entire contents © 2001 by Gartner, Inc. All rights reserved. Reproduction of this publication in any form without prior written permission is forbidden. The information contained herein has been obtained from sources believed to be reliable. Gartner disclaims all warranties as to the accuracy, completeness or adequacy of such information. Gartner shall have no liability for errors, omissions or inadequacies in the information contained herein or for interpretations thereof. The reader assumes sole responsibility for the selection of these materials to achieve its intended results. The opinions expressed herein are subject to change without notice.