Back in 2006, I was blogging on the Radio Userland platform. In that year I was reading a lot and writing a lot of book reviews as blog posts. Today I came upon an old one that I thought I would put on this blog. Enjoy:

The Polaris System Development (Harvey M. Sapolsky)

This book came to my attention at one of NetObjectives‘ recent free siminars.  Alan Shalloway was giving a talk on Lean development and brought a bunch of related books that he let us browse after the seminar was over.  More recently, Dave Smith mentioned it on the AYE Wiki.  I went to Amazon and found that the book was out of print.  It was available used for $229.67 but I can’t pay that much money for a book.  So I sent an email to the Microsoft library.  They didn’t have it but were able to get it on inter-library loan from the University of Washington.  Since reading this book was now time-bound, I paused my reading of Scott Berkun’s book to read this one.

This book is the result of in-depth research into the Polaris System project.  This objective of the Polaris project was to create a submarine-based ballistic missile project (the Fleet Ballistic Missle Program or FBM) in the shortest possible time.  It has been pointed to as an example of an extremely successful project.  It came in early and hundreds of millions of dollars under budget.  Sapolsky was asked to prepare a report on it and obtained permission to publish the report with no review other than for security purposes.  The book was written in 1971, 35 years ago.

The book starts off by setting the context for the project.  It was the late 1950′s and Sputnik had just happened.  America was worried and the various branches of the armed services were all vying to produce an ICBM system.  Soposky does a good job explaining the political manuvering that the leaders of the project had to do in order to get permission to build the system.  Sapolsky talks about four bureaucratic strategies that were used to help make the project happen.  The strategies were:

  1. Differentiation – establishing unchallengable claims on valued resources by distinguishing their program from those of their competitors.
  2. Co-optation – absorbing new elements into its leadership as a means of averting threats to its stability or existence.
  3. Moderation – building long-term support for their program by sacrificing short-term gains.
  4. Managerial Innovation – the attempt of an organization to achieve autonomy through the introduction of management techniques that appear to indicate unique management competence.

The organization of the Navy at the time the project was getting started was composed of 2 groups: the technical bureaus (responsible for material resources) and the operations staff (responsible for determining weapons requirements).  The proponents of the Polaris project set up a special group, the Special Projects Office (SPO) separate from the regular Navy organization that would have complete authority over the project.  Not everyone in the Navy was happy with that idea.  The SPO had to be skillful and make compromises in order to be able to do what it needed to in order to successfully execute on the project goals.  One thing that is stressed in the book is that the bureaucratic strategies that were used did have an effect on the program itself but they were needed to give the program a chance to exist.

Item #4 in the list above (Management Innovation) is very interesting and a whole chapter is devoted to it.  Note the words “appear to” in the description.  New management techniques used in the Polaris project, including the widely-known PERT, were credited with the success of the Polaris project.  So successful was the promotion of PERT and the new management techniques that an industry sprang up to provide PERT tools to all businesses so they could have the benefits of these management innovations.

The trouble was, those innovative management techniques were found not to be responsible for the success of Polaris.  Politically, though, it was in SPO’s interest for the world to believe that the benefit of using PERT was the reason for Polaris’ success.  As Sopolsky writes:

“It was a myth, however, that had value.  For those who wanted the FBM submarines developed, but were reluctant to place their faith completely in the men assigned the task, it gave a sense of assurance.  The management system, not the men, would guarantee the development of the Polaris.  For those who were developing the Polaris, it removed the necessity of justifying each development decision to a higher authority.”

So, what was the real reason for the success of the Polaris project?  According to Sopolsky, careful management of the interfaces of the components of the system and competition for the job of implementing those components.  Sopolsky talks about “maxims” that were used by the top manager for the project.

  • Performance requirements for the system should be set by the technical director of the project along with those in the project who were knowledgeable in the relevant technologies.
  • The program’s objective was the construction of a deployable system and not the advancement of technology. 
  • All technical tasks other than those contributing directly to the deployment of a submarine-launched ballistic missle should be avoided.
  • Naval laboratories were not to be used in the development effort unless they posessed a technical competence unavailable in private organizations.

Focus on high quality was another important part of the success of the effort; “Reliability and methods to achieve it became program fetishes.”

The book continues with an examination of the costs of the program and a look at how the program changed as Polaris drew to a close and its successor, Poseidon ramped up.  The tone isn’t optimistic.

The final chapter reviews the elements that led to success for the project.  These include:

  • A favorable environment – there was immense national support for the mission of the Polaris project.
  • Skill in bureaucratic politics – The leaders of the Polaris were deft in this area and able to protect the project from outside risks in this arena.
  • Ability to manage technical complexity – disciplined flexibility, decentralization, and competition worked extremely well on this project and may be the best way to manage large development projects.

This was an excellent book and well worth the extra effort I had to go through to find a copy.  Reading about this 35-year old project didn’t feel dated at all.  Consider all the recent large development projects you’ve heard about.  Any of them coming in early and under budget?  Maybe they could learn some lessons from the Polaris project.