Book review: “Balancing Agility and Discipline” by Barry Boehm and Richard Turner
Introduction
This book attempts to find the common ground between two differing software development methods: plan-driven and agile-driven. Plan-driven methods are based on strong engineering principles, with an emphasis on a well-defined and -documented process and, usually, a sequential “waterfall” approach of requirements, design, construction and deployment. Examples of plan-driven methods are MIL-STD-498 and ISO/IEC 12207. Agile-driven methods are based on iterative delivery, with a focus on business-value to the customer rather than on the process of delivery itself. Examples of agile-driven methods are XP, DSDM and Scrum. The authors seek to find the middle-ground between the two approaches and describe ways of defining which approach is suitable for which types of project.
Factors
After providing useful contrasts and similarities between agile- and plan-driven approaches, the authors summarise the following as key factors in the decision as to which approach to use in a particular project. These factors are:
Size. Agile approaches discriminate towards small products and teams. Large products and teams will favour a plan-driven approach.
Criticality - in the sense that failure of the system under design causes loss of life and/or large monetary loss. Projects of high criticality will favour a plan-driven approach.
Personnel. Generally speaking, agile approaches require workers that are more highly skilled than plan-driven ones. The rationale is that in the latter, the development process is more highly proscribed and hence easier to learn.
Dynamism. This factor relates to the stability of the project environment; where the business environment is highly dynamic and system requirements are rapidly changing, then agile-driven approach is favoured. In highly-stable environments, the additional overhead of putting in place the infrastructure to deal with rapid change may not be necessary.
Culture. A factor that is often overlooked – maybe the most important one. If the team experience is one where it can adapt quickly to changing circumstances and is empowered to “get the job done” in a turbulent environment, then the agile approach is preferred. Conversely, if the team is more comfortable with clearly-defined job roles and procedures, then the plan-driven approach is favoured (see diagram).
Case Studies
After detailing what a typical projects of each type look like on day-by-day basis, Boehm and Turner turn their attention to two case studies. Interestingly, rather than give a simple case study of each project type, the authors describe projects that migrate toward the middle-ground.
The first relates to a project of fifty staff that started with a pure XP approach but needed the rigour of some plan-driven principles to succeed
The second details a forty-eight-month fixed-price contract to re-engineer a one-million-line missile-detection system. For this project to succeed, it was necessary to introduce more-flexible practices that were recognisably “agile”.
Risk-based Approach
The key part of the book is the authors’ description of a risk-based approach to balance agility and discipline. Some moments’ reflection will reveal the truth in this approach – it is the project sponsor’s attitude to risk (my emphasis) that will determine the balance point. The authors support their approach by describing three fictitious examples:
An event-planning system, where failure would result in loss of venture capital;
A supply-chain management system, where failure would result in a major capital loss;
A national crisis-management system, where failure would result in widespread loss of life.
For each example, the authors assume some reasonable input parameters and then use risk-profiling to derive a development strategy. Interestingly for the crisis-management system, the authors correctly categorise sub-contractors as additional stakeholders – touching on Evolutionary Acquisition techniques.
The final chapter deals with the summary of the findings, namely:
Neither approach is a silver bullet, that is one approach will not be applicable in all situations.
Each approach has project types where one will dominate the other.
In future, a combined approach is more likely to be successful.
Some methods already combine the best of both approaches – DSDM being one of them.
Experience has shown that it is better to start with simple method and build it up rather than start with a heavyweight method and tailor it down.
While methods are important, more critical success factors will revolve around the management of people, value and expectations.
Summary
This excellent book convincingly argues that most real-world software-development projects will incorporate elements of both approaches. The authors correctly identify that a risk-based process is the best way to formulate a method for delivering a software-development project.
Further Reading
"Waltzing with Bears" by Timothy Lister and Tom de Marco. An accessible and insightful book on project risk management. Boehm and Turner’s approach can be extended to use the techniques described here.
"Good to Great" by Jim Collins. A seminal book on the characteristics that turn good companies into world-beating ones. A company wishing to introduce agility to a plan-driven culture (or vice versa) will do well to ensure that their cultural-change programme is aligned with Collin’s findings; as would measurement via a Balanced Scorecard approach (Kaplan and Norton).
Additional Thoughts
At Agilier, we have helped clients reduce costs and achieve timescales by tailoring a method to suit the types of project they deliver and their attitude to risk. Please contact us for a no-obligation discussion.