Tom DeMarco and Timothy Lister sponsored a programming competition in which 166 developers competed on the basis of both quality and speed (DeMarco and Lister 1985). The competitors provided information on the characteristics of their physical work environments, and it turned out that the developers who performed in the top 25 percent had bigger, quieter, and more private offices and had fewer interruptions from people and phone calls than the rest of the group. Table 30-1 contains a summary of the differences in office environments between the best and worst performers.
Table 30-1. Differences in Office Environments Between Best and Worst Performers in a Programming Competition
Environmental Factor | Top 25% | Bottom 25% |
|---|---|---|
Dedicated floor space | 78 square feet | 46 square feet |
Acceptably quiet workspace | 57% yes | 29% yes |
Acceptably private workspace | 62% yes | 19% yes |
Silenceable phone | 52% yes | 10% yes |
Calls can be diverted to voicemail or receptionist | 76% yes | 19% yes |
Frequent needless interruptions | 38% yes | 76% yes |
Workspace makes developers feel appreciated | 57% yes | 29% yes |
| Source: Adapted from Developer Performance and the Effects of the Workplace (DeMarco and Lister 1985). | ||
DeMarco and Lister thought that this correlation between office environment and productivity might be the result of a hidden factor, namely that better developers might naturally have better offices because they had been promoted. But when they further examined their data, they found that that wasn't the case: developers from the same organizations had similar facilities, and their performance still varied.
The data shows a strong correlation between productivity and quality of the workplace. Developers in the top 25 percent had productivity 2.6 times better than developers in the bottom 25 percent. This suggests that moving from a bottom 25-percent environment to a top 25-percent environment is likely to result in a better than 100-percent improvement in productivity.
In the 1970s, IBM studied developer needs, created architectural guidelines, and designed their Santa Teresa facility with developers in mind. Developers participated throughout the design process. The result is that the physical facilities at Santa Teresa are rated the highest in the company each year in annual opinion surveys. IBM reports that the facility improved productivity by about 11 percent—and IBM didn't start out in the bottom 25 percent (Jones 1994).
Office space in my area costs anywhere from $1.00 to $2.00 per square foot per month. Developer time, including salary and benefits, costs anywhere from $4,000 to $10,000 per month. The average cost to go from the bottom 25 percent to the top 25 percent (46 to 78 square feet and least-expensive to most-expensive office space) would be about $110 a month per developer. The average productivity improvement is calculated by multiplying productivity by a factor of 2.6, which would work out to about $11,000 a month.
A decision to skimp on office space seems to me to be "office-space-wise and developer-time-foolish"—by a factor of about 100.
All in all, an organization that currently has an average environment can probably expect to improve its development productivity by 20 percent or more by providing its developers with productivity environments. The gain for organizations that currently have above average or below average facilities will vary accordingly.