The work of John Boyd is becoming more well known in the IT community. However, this has, so far, been based on an oversimplified view of Boyd's theories - notably his so-called OODA loop. This is unfortunate as both Boyd's OODA loop and its wider theoretical underpinning potentially offer valuable insight into why IT systems often fail to meet the ongoing challenges of a fluid operating environment.
 |
| John Boyd (1927-1997) |
Who was John Boyd?
John Boyd was a military strategist. As a young fighter pilot Boyd was fascinated with the reasons why some planes and some pilots were successful in dogfights and others were not. In order to make meaningful comparisons between fighter planes Boyd created Energy-Maneuverability (EM) theory. EM theory uses mathematical models of the behaviour and capabilities of different aircraft to compare how each would fare against the other in various circumstances. From the analysis of both US and Soviet fighters Boyd drew two conclusions: Firstly that in many cases Soviet fighters were superior to US fighters and secondly that the metrics which were then used when designing and procuring aircraft were not overly relevant to actual air combat. At the time of EM theory fighter aircraft design was driven by absolute required maximum speed of the plane together with its ability to carry huge amounts of ordinance. Boyd’s EM theory showed that in any likely combat encounter the winning plane would likely be the one which could change its state most quickly. For that reason he concluded that fighter planes should be as light, powerful and manoeuvrable as possible. These planes should also be cheap enough to buy in large numbers. This subsequently led to the development of the F16 which was both cheaper than its predecessors and also more manoeuvrable.
Whilst EM theory was focused on understanding the comparative dogfighting merits of fighter planes Boyd's other important theory - the OODA loop - was concerned with how individuals and groups could understand their changing environment and subsequently make timely and appropriate decisions regarding how to respond to it.
The OODA loop has four main phases:
Observe - gather information on the situation around you.
Orient - build a model of the world using the information gathered together with your cultural and genetic heritage.
Decide - plan how to react to the external situation.
Act - execute the plan.
Iterate. Iterate. Iterate. The quicker you can iterate through the loop the quicker you can react to your environment and, in combat, the more likely you are to defeat your enemy.
The 'OODA loop' is often characterised (i.e. misrepresented) as a very simple cycle. However, Boyd's actual OODA is much more complex as shown below:
The misinterpretation of Boyd's work (especially the OODA loop) is understandable since he wrote little of it down in a formal manner. Aside from a few presentations and his earlier work on dogfighting tactics and modelling of aircraft combat performance the only written work which exists is his nine page paper Destruction and Creation. Destruction and Creation was Boyd's attempt to understand and document the mechanisms by which creativity can occur in organisations - an important concern for Boyd as much of his war fighting strategy was based on the hypothesis that the winning side is the one which better maintains an overall understanding of the current situation and which can therefore operate at a higher tempo thus not only being more 'agile' but also degrading its enemies' model of what's going on.
Boyd and IT
For some reason, probably its perceived virile masculinity, military strategy has long been influential in business strategy. Sun Tzu is often read by business people who see themselves as warriors on the battlefield of commerce. This partly explains the attraction of Boyd to some. Also, from both a business perspective in general and an IT perspective in particular Boyd's OODA loop is attractive as it seems to offer what many are seeking; agility. In IT the Agile method is becoming the default mechanism for software development throughout the industry. Who would deny that businesses or IT development processes should demonstrate agility?
However, I would argue that in focussing on a (highly simplified) version of the OODA loop much of its value is lost. Additionally, the intellectual and attitudinal underpinnings of the OODA loop are at odds with some of the ways in which people are attempting to implement it in IT.
So, how should we apply the OODA loop?
The OODA loop can be applied as a general decision making model and as a mechanism for ensuring that an organisation is able to continuously adapt to a changing competitive environment. Clearly, Boyd did not see the adoption of his loop in terms of a fixed set of tools and processes. Since Boyd viewed people and mission as being more important than tools (weapons) it is likely that he would place an emphasis on ensuring that all leaders (at all levels) in an organisation shared a common set of attitudes, approaches and cultural values together with a common understanding of the organisation's mission (Schwerpunkt in Boyd-speak). Additionally, people in leadership positions should be sufficiently creative so as to be able to operate effectively with a high degree of operational autonomy.
Clearly business is not a zero sum game in the same sense as a dogfight - i.e. for a business to succeed it is not necessarily required for it to destroy its competitors. Therefore at least some of the point of the OODA loop in conflict is (at worst) not required and (at best) a helpful side-effect. Specifically one of the effects of the OODA loop when applied in conflict is to create chaos and channel it in the direction of one's opponent. This destroys the opponent's model of the rapidly developing situation and stretches his or her decision-making cycle. In business it is clearly advantageous to operate at a faster tempo than your competitors and also to behave in ways which do not fit their models of the world.
Beyond the facile interpretation of the OODA loop as a mechanism for making decisions quickly is Boyd's belief that the key to an organism's survival is its model of the world, its relevance to reality and the synthesis which it performs on this analysis of the world. In Destruction and Creation Boyd attempted to describe how creative thought based on analysis and synthesis can lead to innovation. He proposed a thought experiment in which four domains were to be considered; a skier on a slope, a speedboat, a bicycle and a toy tank. Each of these domains had its own attributes: chair lifts, skis, people, mountains, chalets for "skier", sun, boat, outboard motor, water skier and water for "speed boat", chain, seat, sidewalk, handlebars, child and wheels for "bicycle" and, finally, turret, boy, tank treads, green paint, toy store and cannon for "toy tank". Each of these domains is therefore associated with a network of interlinked concepts. Boyd argued that it was by breaking these links, dissociating them from their domains, and then reassembling them in novel ways that new ideas were formed. Boyd illustrated this by picking several of the previously identified attributes; handle bars, outboard motor, tank treads and skis. Combining these allows us to create what Boyd termed a "new reality" - or in this case a snowmobile. Boyd wanted the US military to be a culture which encouraged its officers to be the creators of snowmobiles - to be individuals who were capable of the analysis and synthesis required to think creatively. Boyd went on to argue that having created a snowmobile (or any other concept) it must be measured against external reality and (even if initially) a good fit subsequently itself destroyed and its conceptual network rebuilt to fit a changing external environment. This cycle of destruction and creation underpins Boyd's OODA loop. Boyd referred to the works of Gödel, Heisenberg and the second law of thermodynamics in order to support his theories. Using Gödel's Incompleteness theorem Boyd argued that any system representing reality must necessarily be inaccurate and incomplete. Using Heisenberg's Uncertainty Principle Boyd argued that in combat the observer effect means that the neat distinction between protagonists becomes blurred. The Second Law shows that the entropy in any natural system will increase thus making attempts to understand it increasingly difficult.
Where are the lessons for IT here?
There are two main areas where Boyd's work is applicable to IT (and to an extent business itself); organisational culture and concept formation.
Organisational Culture
Underpinning Boyd’s belief that manoeuvre warfare was preferable to attrition warfare were his views on the levels of autonomy and creativity which should be displayed by officers at all levels in the hierarchy. In order to successfully adopt manoeuvre warfare an army will need to delegate decision-making which is made at the top of the hierarchy down to commanders in the field. In order for this to be viable the officers involved need to share both a culture and a common understanding of the mission involved. Therefore even when these officers are exercising a great deal of operational autonomy they are doing so in ways which mesh with others and also effectively serve the wider mission. This approach enables an army to simultaneously execute multiple OODA loops at different organisational levels.
Does this approach work for IT? Well, it clearly offers merit from a purely business perspective. Organisations which adopt a top-down 'attrition-warfare' approach to decision-making are clearly less able to react to a rapidly hanging environment. However from an IT perspective the situation is slightly more uncertain. The OODA loop is designed to allow a protagonist to create chaos and then channel it at his or her enemy. Is the creation of chaos the purpose of IT? Despite much evidence to the contrary it clearly is not. IT is, fundamentally, concerned with creating useful models of the world. These models get realised as IT systems - executable code, databases, servers, networks etc. So if an organisation were merely to adopt a Boydian approach to delegation of authority for IT then we could expect a period of rapid change in IT systems followed by a period of head scratching as the organisation dealt with the ensuing complexity which had been created - or 'accidental architecture' in IT parlance. The term 'software' is actually a little misleading. Software isn't soft. Once implemented software is actually relatively expensive and time-consuming to change. So, for a Boyd-like approach to work it needs lower-level decision makers to operate within a meaningful governance framework which provides a structure in which local decisions can be made without compromising higher-level or longer term goals and strategies. In IT architectural terms this would equate to a lightweight implementation of an architectural governance approach/framework such as TOGAF. This clearly needs to be done in such a way that the correct balance between top-down governance and local agility is struck.
The second implication for the structure of organisations concerns the people themselves. Boyd argued that the officers in a manoeuvre warfare army should be 'Snowmobile builders' - i.e. Individuals who were capable of both analysis and synthesis. To some extent creativity can be taught or at least fostered. However, organisations wishing to implement Boyd's ideas would be advised to hire individuals capable of demonstrating creativity - in this case by being able to disassemble existing concepts and the reassemble in new and useful ways.
Concept formation
The second major lesson for IT from Boyd is around how concepts are formed, destroyed and reformed. During the orientation phase (or 'The Big O' as Boyd termed it) of the OODA loop the analysis of gathered data and the synthesis of a new or modified view of the world takes place. In the heat of battle this model is used to quickly generate plans which can be reused against the enemy. However, in commerce these models get encapsulated as IT systems. Once so encapsulated they are relatively hard to change. This prevents organisations from rapidly iterating an OODA loop. The entropy inherent in the external environment gradually destroys the value of legacy IT systems. As in manoeuvre warfare the ability to be creative and to react to a situation by modifying our internal model of the world is critical in a commercial environment. Additionally, philosophically, Boyd's theories serve to illustrate that models are inherently a simplified, incomplete and inaccurate mechanism for understanding the world and they only get more wrong over time.
What is slightly unfortunate is that this lesson from Boyd really only serves to highlight the problems and imperatives of which IT folks are already painfully aware - i.e. the need for agility and the barrier which complicated legacy systems present. We're still developing the tools both in terms of financial models which describe the ongoing and opportunity costs of complexity and also in terms of software design patterns and practices to address complex systems.
Perhaps what we really need is an IT version of Energy-Manoeverability theory which instead of quantifying the flight characteristics of warplanes instead helps us to understand the qualities and characteristics of computer systems - ideally along with the associated costs.
For anyone with an interest in Boyd's life and theories the books "Boyd: The Fighter Pilot Who Changed The Art of War" by Robert Coram and "The Mind of War: John Boyd and American Security" by Grant Hammond are both highly recommended. The former focuses on his life and the latter focuses more on his work and theories. Additionally, his original briefings are all available at Air Power Australia
