It took me some time to warm to this book. Nothing much happens in the initial 80 pages. The first chapter develops two fairly tortuous case studies - the vicissitudes of fortune in the recording industry in the last decade and the struggle of the Apaches against the Spanish invaders - to introduce the theme of the book. Then follows a discussion of the morphology of decentralised organisations (in terms of power distribution, funding, etc). Chapter 3 illustrates these formal characteristics with a series of examples, ranging from Skype over Wikipedia to Burning Man. There is honestly not a lot of meat to chew on in these first chapters and some patience is required from the reader.
It becomes more interesting in Chapter 4 where Brafman and Beckstrom discuss operational principles behind decentralised organisations (the need for pre-existing networks as a substrate, the role of catalysts and champions to activate leaderless organisation, "circles" as their chief co-ordination mechanism, and "ideology" as the glue holding everything more or less together). The role of the catalyst as a "servant leader" (term, however, not used by the authors) is further elaborated in the fifth chapter.
In chapter 6, the discussion turns to the question "What do you do, as an incumbent, when you are under fire from a starfish?" It transpires that there is not an awful lot to be done: you can try to morph them into a spider by activating internal cancer cells (greed and competition), you can try to dissolve or change the glue, the ideology that keeps the structure together or you can join them and become decentralised too (then it's starfish against starfish).
Brafman and Beckstrom maintain that it is not always necessary to go all the way and radically decentralise. There is such thing as a "hybrid" organisation (Chapter 7), which mixes principles of centralisation and decentralisation. Here the discussion suddenly gets denser and this is a part of the book that warrants repeated reading. A distinction is made between centralised organisations that give customers a voice (eBay with its peer-to-peer feedback is an example), those that put their customers to work (IBM developing open source applications) and those that decentralise parts of their internal structure. Towards the end of the chapter, however, the discussion peters out. "Appreciative Enquiry" is invoked as an approach to bring a whiff of decentralisation into companies who want to hang on to their centralised bureaucracies. It's a dangerous example that may tempt people into crass opportunism (that is, however, bound to backfire on them).
Finally, the authors hypothesise that in a given ecosystem there is no static equilibrium in terms of right mix of centralised/decentralised characteristics ("right" in terms of securing survival and the ability to extract economic rent). The "sweet spot" changes as a function of time, sometimes dramatically so. The desire for anonymity and the free flow of information are forces that push towards the decentralisation end, whilst the desire for security and accountability pull the system back to a more centralised mode of operation.
The book closes with a short epilogue that lists 10 simple guiding principles to make the most out of decentralised organisations or to defend yourself from their attacks.
On the whole, I enjoyed this book. It provides an intelligent and accessible discussion of a complex issue. With respect to the latter, the authors do a laudable job in keeping thing simple, but sometimes it's over the top. Particularly in the first halve of the book, their penchant for telling anecdotes and stories makes them err on the side of the trivial (a discussion on Wikipedia starts with "we all remember doing school reports in the sixth grade. Back then, research meant going to the library and hoping the that the Encyclopaedia Brittanica wasn't checked out ... and so on, and so on.) I was irked more than once by the patronising and befuddling prose of Brafman & Beckstrom. Admittedly, sometimes they hit it right. The title of the book, for example, is a very strong and aptly chosen metaphor for decentralised and centralised organisations, respectively.
Also I believe this book does not exhaust the potential of this fascinating subject matter. I think the discussion would have gained significantly in clarity and power if only a number of well known systems science principles (such as Ashby's Law of Requisity Variety, see
Introduction to Cybernetics (University Paperbacks)) had been invoked to give the whole discussion a rock solid footing. I also missed a solid link to the burgeoning literature on the P2P movement. It is clear that the issue of property rights in central in making leaderless organisations work (Brafman discusses this as a way to sabotage starfish only) and people like Lawrence Lessig ("
Free Culture: The Nature and Future of Creativity) and Yochai Benkler ("
The Wealth of Networks: How Social Production Transforms Markets and Freedom") have a lot to say about these issues.
A small point, but a fairly irritating one, is the use of the word "ideology" in the book. The authors ostensibly use this to refer to any set of beliefs that underpin a decentralised organisation. From my point of view, the word "ideology" refers to a more elaborate and closed system of abstract thought (and as such has a pejorative tinge to it). Many starfish (also amongst those mentioned in the book) thrive on a much more vague and fluid set of beliefs, norms and values. It's worthwhile to be more nuanced about this.
Morally speaking, the book leaves the reader in suspension. From an internal point of view, leaderless organisations are unquestionably superior - morally and aesthetically - to centralised organisations, not only because of their structural simplicity and elegance, but also because they rely so openly on trust (in my opinion THE key word in the book), on the belief that man is fundamentally good and ultimately because they are capable of drawing the best from people and providing them with truthfulness, meaning and purpose in their life. Problem is that not only Alcoholics Anonymous operates as a decentralised organisation, but Al Qaeda does too. So starfish can server all kinds of purposes, some more constructive than others. It all depends which side you're on.
