Why do we need Management of Software Development?
*Adapted from David Root (2014)
Why do so many projects fail?
CHAOS Report 2015 - Project Success Rates by Size:
| Project Size | Successful | Challenged | Failed | Total |
|---|---|---|---|---|
| Grand | 6% | 51% | 43% | 100% |
| Large | 11% | 59% | 30% | 100% |
| Medium | 12% | 62% | 26% | 100% |
| Moderate | 24% | 64% | 12% | 100% |
| Small | 61% | 32% | 7% | 100% |
The Standish Group CHAOS report (2015) shows a clear correlation between project size and failure rates. Larger projects face significantly worse outcomes:
- Grand projects: Only 6% succeed, with 51% challenged and 43% failed
- Small projects: 61% succeed, with 32% challenged and only 7% failed
Overall across all project sizes, approximately 36% of projects are successful, 45% are challenged, and 19% fail.
Success: Delivered on time, within budget, and with the required features
Failure: Cancelled before completion, or delivered but never used
Challenged: Late, over budget, and/or missing required functions
Despite decades of experience:
- CMU CS department is 60 years old
- UC Berkeley’s is 40 years old
- NASA has used software since the 1950s
Yet, software development remains unpredictable compared to other engineering disciplines.
Few Thoughts
Software is Different
- Orders of magnitude complexity: Software systems can be vastly more complex than physical systems.
- Intangible: You can’t touch or see software; it’s abstract.
- Perceived as “soft”: Easy to change, but changes can have unexpected consequences.
- Abstraction: Both the product and the process are abstract.
- Stakeholder understanding: Many stakeholders do not fully understand software or its development.
- Uniqueness: Every software project is different; there are few universal solutions.
Example Question:
Can code written in BASIC and in C# compile to the same machine code?
Answer: Yes, if both are compiled for the same architecture, but the process and efficiency may differ greatly.
The Human Variable
- #1 problem for managers: How to motivate people?
- Standardization challenges:
- Experience varies
- Abilities to learn and execute tasks differ
- Interpersonal relationships affect team dynamics
- Unpredictability: Human behavior is hard to predict and control.
- First fit problem: Matching people to tasks is rarely perfect.
Constraining the Human Variable
- Processes:
- Define acceptable behavior norms
- Direct behavior and actions
- Standardize responses: “See A, do B” (first fit)
- Pitfall: Over-standardization can lead to “assembly line” problems—loss of creativity and adaptability.
- Process improvement: Continuously refine how work is done.
- Process checking: Ensure processes are followed and effective.
- Communication: Historically, communication in software teams is poor, leading to misunderstandings and mistakes.
Why Study Management of Software Development?
- To understand why projects fail and how to improve success rates.
- To learn how to manage complexity and human factors.
- To apply engineering discipline to software, while recognizing its unique challenges.
- To develop skills in process design, team leadership, and effective communication.
Acknowledgments
This content is heavily inspired by and adapted from lectures by Eduardo Miranda and David Root on software project management. The structure, examples, and pedagogical approach reflect their teaching materials and frameworks.
Sources
- Root, David. Managing Software Development. Lecture materials, 2014.
Disclaimer: AI is used for text polishing and explaining. Authors have verified all facts and claims. In case of an error, feel free to file an issue.