Risk management through the lens of complexity-- the BP oil crisis



Complexity theory is based on the philosophy that total order does not allow the flexibility necessary to address human situations that happen in business dynamics. Western society is moving to the concept that less order allows greater flexibility while addressing the concern that less order leads to less control. This paper examines management through complexity theory of the British Petroleum (BP) oil spill in the Gulf of Mexico. Although there is much to learn from this crisis, the results proved to be a victory for complexity theory. This paper relies upon current information, including the findings of the National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling appointed by the President Barack Obama to investigate this incident, as well as the internal lessons learned official investigation conducted by BP. The findings of the research show how complexity was deployed in response to the unprecedented buildup of people and equipment, and how management at BP had to evolve into dynamic complexity based management rather than relying upon its published linear risk plan.


BP staff and management addressed the Gulf oil spill in the traditional manner outlined by the company's published spill response plan. Once the crisis happened, BP moved forward with an organized response as outlined by the plan. What developed was that the plan was underdeveloped to address a situation of this magnitude and the situation quickly overtook the individuals involved. The process showed that initially the response seemed to follow a normal, formal risk plan, but it was improbable such a system would be successful. Once the magnitude of the spill was accepted, BP and other organizations needed to mobilize a vast armada of equipment, people, and materials to combat the spill. The spill was larger and more complex than anything that had been seen prior in the United States and so it became necessary to develop new systems and processes in order to be successful in the shortened amount of time.

These are the four critical findings discovered regarding the BP response to the Gulf oil spill:

  1. The amount of people and equipment involved in the process was an unprecedented build up in an amazingly short period of time, a situation that was not covered in the existing BP response plan. This required a complexity based system to mobilize and de-mobilize the vast resources necessary to combat the oil spill.
  2. Traditional risk planning was abandoned in favor of on the spot response where management was able to respond dynamically as the BP spill response plan was very poorly conceived and written and unable to cope with the magnitude of the Gulf spill.
  3. BP now acknowledges the need to move the center of operations from a static base in Houston, TX, (as outlined in its response plan) to an on the spot management empowered to make rapid decisions as new information became known.
  4. Communications locally were identified as critical to the success of the spill cleanup.

In support of these findings, a review of the formal findings of the situation has been done. A review of open source literature, including newspaper articles, journal articles, trade journal articles, blogs, and other sources, were examined to understand the circumstances surrounding the incident. There has been considerable material written regarding the incident, however, much of the material available focuses on the sensational aspects of the event rather than upon the actual learning from the event. In this regard, much of the available material needed to be excluded or carefully selected to maintain the focus of this paper. Although the authors agree that the environmental impact of this catastrophe is considerable, the goal of this paper is to find new learning in order to mitigate the impact of such a calamity in the future.

Risk Management Through the Lens of Complexity – The BP Oil Crisis

Amid all the uncertainties will be opportunities. Otherwise untouchable oil assets are likely to be sold to raise billions to pay the price of the historic spill in the Gulf of Mexico. The pressure to rapidly apply the lessons from the disaster in the Gulf will benefit those who develop, or acquire, next generation innovations able to better manage the risks revealed. (Deloitte, 2010, p. 3)

Complexity to Support the Buildup on People, Materials, and Equipment

There is sufficient data to support the buildup of people, material, and equipment in a very short time. Part of this buildup has been from the poor initial estimates of the magnitude of the spill. Early estimates were showing that the maximum amount of the spill was 5,000 barrels (National Commission, 2011), which was shown to be a completely inaccurate forecast. This initial forecast hampered the increase of the response because the BP spill response plan requires that an estimate be used as a first step (BP, 2010). These rigid elements hindered the initial attempts in containing the spill. The estimates were entrenched as fundamentally accurate, which further caused delays in the deployment of additional equipment and material.

Since the original high estimate was 5,000 barrels per day (which was externally confirmed outside of BP), BP responded and deployed dispersant material to handle the 5,000 barrel a day spill. When the initial deployment of dispersant was found to be inadequate, the only two possibilities were that the dispersant was defective or there was substantially more oil in the water as to make the quantity of dispersant ineffective. Precious time was lost because there was the feeling that the estimate was correct and so time was wasted as the dispersant was being blamed as being defective. Once the dispersant was found to be in good working order, the only possible alternative was that the 5,000-barrel estimate was inaccurate.

At this juncture, BP should have considered both possible options instead of pinning hopes on the dispersant defect. The decision to look linearly at the problem, rather than work on multiple possibilities shows how the limited risk plan hampered the initial attempts at spill containment. Early failures like this helped BP move from following the entrenched risk plan to a more dynamic method of addressing the issue. The sad point is that it took so long for BP to realize this deficiency and only later did they accept the magnitude of the spill and moved to respond appropriately.

“Complex systems almost always fail in complex ways” (National Commission, 2011, p. vii). In the end, the risk plan developed by BP was fundamentally flawed (if one were to ignore the obvious glaring mistakes, such as information on how to handle sea lions and walruses in the spill plan, and the fact that the environmental consultant identified was dead years before the spill plan was submitted) in that the plan spends too much time discussing a very linear approach to an oil spill (National Commission, 2011).

The plan includes pages of flowcharts on how and when to deploy dispersants for an oil spill. Furthermore, the deployment of dispersant is only considered if the oil is heading towards shore or colonies of sea birds. No other wildlife is considered as part of this process. If the flowchart allows for dispersants to be used, there is an internal requirement where approval must be sought in order to apply dispersants. This linear requirement seems to ignore that on the spot action might serve better than a hierarchical and linear response. Only after the spill is flowing out of control and national attention has been attained does BP start to move with more alacrity.

It is interesting to note that even after BP submitted paperwork as the responsible party on 24 April 2010 that the mobilized response was quite small. On 28 April 2010 when the U.S. Coast Guard identified BP as the responsible party, the mobilization effort begins in earnest. Within two days, BP moved from 500 people deployed (mostly for call centers to handle claims) to 2,000 people and 75 vessels were deployed on April 30, 2010. It is clear that a shift occurred within BP management that the Gulf spill was no longer a simple matter of closing the well, applying dispersants, and deploying skimmers. The leak was larger than anyone had ever anticipated and the process to clean up the spill and to close the well would take months. No simple risk plan flowchart would be able to cover a situation of this magnitude.

Abandonment of Traditional Risk Management

To understand the order of magnitude of the spill and the escalation of the response, Exhibit 1 shows the rapid buildup of people and vessels that were involved in the clear up efforts. In addition, BP also deployed anywhere from 24 to 120 aircraft to assist with spotting spills in order to direct the vessels towards the oil.

Rapid buildup of people and vessels involved in the cleanup efforts (British Petroleum, 2011)

Exhibit 1: Rapid buildup of people and vessels involved in the cleanup efforts (British Petroleum, 2011)

Consider Exhibit 1 and understand the magnitude of this situation. BP had a limited presence in the U.S. Gulf region, mostly surrounding the exploration of oil in the region to being the single largest employer in the region having up to 46,000 people involved with spill response. Not only did it have this large human resource, but BP also had to contract and charter over 5,000 vessels and over 100 aircraft at the peak of the incident (British Petroleum, 2011). Consider how large of a task this would be to be able to flex to a point that an organization with a limited presence would go to engaging 46,000 people all focused upon a single task—the containment of an oil spill.

In order to understand the magnitude of the number of vessels being deployed, consider that according to 2009 estimates, China has approximately 300 vessels in its navy and Australia has 51 vessels in the Royal Navy (Measuring the Chinese fleet, n.d.; Royal Australian Navy, n.d.). This means that BP had to move from no ability to manage a fleet of vessels to be able to mobilize, outfit, and organize a fleet of over 5,000 vessels. As the example shows, nations have considerably more time, tradition, and resources to organize a surface fleet, while BP had to find a way to command a fleet of over 5,000 vessels.

Furthermore, consider the demobilization requirements necessary to address this change. Clearly, a complexity based system needed to be developed in order to address these requirements. There is no doubt that a traditional system of hiring and chartering needed to be abandoned in order to support such a huge buildup (British Petroleum, 2010).

Very quickly in the process, the risk management system of BP was overwhelmed by the multitude of requirement being imposed by the national, state, local governments, coupled with the negative press and ill feeling generated in the area. BP became vilified for the spill and the company was seen as uncaring to the plight of those affected by the spill. In order to combat this mutiheaded public relations disaster, it required a complex and dynamic system to address all the challenges encountered. BP not only had to address the media but they also had to spearhead the capping of the spill as well as coordinate the environmental relief efforts associated with the spill.

Example of How Complexity Was Deployed to Replace Traditional Risk Management

BP employed complexity because BP was forced to show results before the project team might be ready. This was seen with the media coverage of the Top Kill project. Top Kill was also unsuccessful with another spill but the impatience of public opinion forced this project to move forward. Any veteran project manager will agree that stakeholder impatience and haste can often create the necessity for non-sequential activities. Linear project managers might find themselves paralyzed by this need to show results. In this case, the project manager (BP) offered excuses that represented reality but the need for action was seen as more important than the more likely solution, drilling a relief well.

Project managers can be pushed to resolve and handle issues out of the typical sequence in order to achieve certain milestones that are important at a higher level (Weaver, 2007). This kind of pressure can be exerted upon a project manager in order to achieve certain milestones faster in order to achieve quicker results. BP fell into this trap and forced into a situation where inputs lead to outputs (results) (Exhibit 2).

Inputs lead to outputs (Curlee & Gordon, 2010)

Exhibit 2: Inputs lead to outputs (Curlee & Gordon, 2010)

BP eventually used a complexity-based strategy by starting a campaign to address the negative press. BP started advertising its organizational values to better explain how the situation would be addressed. BP began leveraging its network in order to address this troubled project. There is no doubt this is a difficult decision for the organization as it does expose them to potential short term failure, but if the organization is expecting society to offer open and honest communication then BP must be guided by that same value (Duarte & Snyder, 2006). The result of not adhering to the value of open communication is far worse than any smaller issue, which may arise from this type of problem. If one expects to use complexity to assist with these types of systemic and social challenges one must be certain the organization, in this case BP, must continue to model and share the values of the organization.

Local Deployment of an Incident Control Post

To support this massive buildup and scale back, it was necessary for BP to establish an Incident Command Post (ICP) in Houma, Louisiana. As the situation evolved, additional ICPs were established in order to manage the scale and uncertainty (complexity) of the spill. Over time, the response teams altered the established response structure to ensure support and authority flowed rapidly to local leadership and communities. This was critical as it allowed leveraging the local community in the cleanup efforts. A complex and comprehensive view of the location of the surface oil, coupled with robust communication among leadership, airborne units, and responders on the water, became the most effective way to deploy the necessary resources. As the spill continued, there was considerable concern regarding hurricane season approaching, the local team needed to flex its long-term plan to address issues such as boom and skimmer supply and placement and inclement-weather operations (British Petroleum, 2010).

Considering what needed to be done, it is not surprising that BP would give up centralized operations in Houston for a local system to support the local efforts. Since communication and location identification of deployed vessels was not easy to coordinate, it was better to have a local command post to address the local situation (British Petroleum, 2011). In addition, allowing for a local command post with local support and resources was better accepted than having some distant operations dictate decisions. It is conceivable with ample communications and vessel location identification that a central location such as Houston might have been effective. The issue was that having a distant central authority gave the appearance of a lack of addressing the local issues at hand.

Consider that finding the oil spills was difficult, and in turn deploying the right vessels to address the situation became equally challenging. Then contemplate balancing supply chain and human resources issues that come with any large scale, unplanned deployment (British Petroleum, 2011; British Petroleum, 2010). Trying to control such a monster at a great distance would be considerably difficult, under the best of circumstances, and it would be completely impossible, in any kind of unexpected circumstance. Departing from a linear and hierarchical system in favor of a local management structure empowered to make quick decisions became a key learning and necessary element in the future.

BP spill response effectiveness (as measured by barrels skimmed/burned) (British Petroleum, 2011)

Exhibit 3: BP spill response effectiveness (as measured by barrels skimmed/burned) (British Petroleum, 2011)

Exhibit 3 illustrates how the response escalates over time. The initial response, hampered by poor information about the spill and the lack luster response by BP are clear by the lack of oil skimmed or burned. Exhibit 3 shows that clearly the stride of the response efforts came about 30 days after the initial sinking of the Deepwater Horizon. One can see the relationship between the change to a distributed, local management and the spill response effectiveness. As BP responded locally, the efforts became more focused and directed. A dispersed management team that was focused locally became the most effective management methodology to combat the spill.

The New Communication Model at BP

One of the crucial learning was that in order to be effective in locating and cleaning up the oil, there had to be an effective manner to coordinate the vast amount of equipment and resources necessary to combat the spill. Since spill containment requires a variety of tasks (application of dispersant, booming, locating oil, moving equipment to the located oil), a local containment team needed to be present and in continuous communication. Also, the support fleet surrounding the spill site required coordination of drilling rigs, multipurpose support vessels from fireboats to tugs, all operating in close proximity under hazardous conditions due to the presence of volatile hydrocarbons in the water.

Since all prior BP operations have only involved three or four ships in proximity of a platform, considerable planning had gone into any time BP entered into a situation that required simultaneous operations. Since vessels would have to operate within 40 feet of one another, the positioning of vessels was deemed critical and all movements were planned out in advance.

In order to cap the Mancondo spill, the ultimate operation required the simultaneous operation of up to 19 major vessels, each up to 825 feet in length within a 1,650-foot radius of the wellhead, and up to another 50 vessels operating in a one-mile radius. If this were not complicated enough, there would be times that vessels would be operating within 25 feet of one another. Given the complex nature of vessel operations, it was critical for vessels to continually understand each other's location as well as to carefully coordinate all of the movements and actions of these vessels (British Petroleum, 2011).

Since planned movements were not possible in advance, this complex operation required a new system that would allow the dynamic movement of vessels, in a safe and expedient manner. One of the improvements that assisted in the simultaneous movement of equipment was the establishment of a rotating on-site branch director. The on-site director was responsible for the 24/7 operations and worked in coordination with the Houston-based team. Simultaneous operation also leverages the continuous use of storyboarding to allow team members to visualize the precise positioning and maneuvering of vessels. This process is not new and is commonly used in military operations where generals and admirals are given a better understanding of military operations by being able to visualize the movement of equipment.

A technology innovation that was identified as important to the simultaneous operation was the automatic identification software (AIS), where an on-site transponder is place on vessels in order to allow for individuals involved having real-time visualization, identification, tracking, and positioning of vessels on graphical displays (British Petroleum, 2010). Many of the vessels in the incident area were equipped with this relatively new technology, however, not all. BP clearly identified that need for this type of equipment to be installed on all vessels operating in a situation like this, as this type of technology would support the relief efforts. This improvement in identification and tracking allows operations to evolve without preplanning as those in command of the operations would be able to move about vessels operating in close proximity with a degree of certainty and safety. This allows for vessels that are operating in difficult conditions to work together with less risk. This is a more complexity-based solution where vessels are offering positioning information rather than the traditional and hierarchical method where the vessel would radio in its position and its movements.

Risk Management and Complexity

Project managers must deal with risks and opportunities that will continually have to be planned throughout the lifecycle of the project. In addition, the project manager will have to provide the team with the tools and confidence to react to the unexpected in order to be successful in changing circumstances (Weaver, 2007; Pritchard, 2005). The Gulf oil spill was a project of an order of magnitude not seen before. Although BP started with a linear approach, it became very clear that such rigid restrictions were not in the best interest of the project. The Gulf spill became a project that quickly attained global significance along with a need to show rapid results. Complexity theory recognizes that such complex projects require an open system, such as an anthill, that can react to the unexpected quickly and revert back to the norm or to a new norm. An open system like an anthill is more likely to survive long term as the open system in fundamentally better at adapting to new circumstances.

What was seen was that the project needed to integrate the unexpected into the project and that the role of the leader is to provide guidance in order to have the project completed in the shortest possible time. The project manager has to balance risk management between the traditional risk management process and preparing the team to expect the unexpected. Project managers must not believe that by thoroughly following a project risk management plan that the unexpected will not happen.

As a result, BP has updated its risk management plan that incorporates these elements of complexity theory:

1. The implementation of robust, proven systems and tools for planning and implementing the management of large numbers of vessels at extremely close quarters, including storyboarding and a centralized, onsite control regime (British Petroleum, 2010).

This dynamic and continuous communication is designed to avoid a static hierarchical system of communication. Instead of relying upon a hierarchy and chain of command, a complexity based network communication is necessary for success (Curlee & Gordon, 2010).

2. The deployment of AIS as an enabling technology for real-time visualization and management of offshore marine operations (British Petroleum, 2010).

The need for dynamic technology to allow real time visualization and managed of the offshore fleet is critical to success. This is another manifestation of a network communication where vessels are communicating automatically position, speed, and location without any hierarchical or linear communication.

3. Demonstrated protocols for directing vessel traffic in the presence of flaring, even with the continuous incidence of VOCs and the need to ensure that levels were below the lower explosive limits (LELs), as well as new techniques for managing the presence of these hazards (British Petroleum, 2010).

Command moves from being solely an order driven organization, to one that looks holistically at the efforts. Command can offer direction; for not only traffic, but also it can offer information that can assist in vessels avoiding hazards that are in the area. This global perspective is one that is required in a complexity driven organization.


Complexity is everywhere and more than ever complexity can assist business in achieving greater results with fewer resources. The more that manager's can apply complexity, the better that they will be able to manage others. Complexity is about understanding the small in a way that it can be applied to the large (Curlee & Gordon, 2010). The BP spill has shown that projects will become larger and more complex and the project managers that learn to leverage complexity will be able to handle larger mega-projects or programs (McKinnie, 2007). Project management, as a whole, needs to recognize that complexity theory is now a force in project management. Despite that A Guide to the Project Management Body of Knowledge (PMBOK® Guide)—Fourth Edition (Project Management Institute, 2008) is currently lacking information about complexity theory, there are projects, including the efforts by BP that have successful deployed complexity theory. Currently, there are many competing leadership techniques that are available to project managers; however, none offer the same applicability and flexibility as complexity. The Gulf oil spill disaster has shown that even large organizations can learn to successfully apply complexity theory. The more that organizations can learn from their mistakes, the more successful that they will be with the changing global environment.

BP oil spill response plan. (2010). Retrieved from http://info.publicintelligence.net/BPGoMspillresponseplan.pdf

British Petroleum. (2010, September). Deepwater Horizon containment and response: Harnessing capabilities and lessons learned. Retrieved from http://www.bp.com/liveassets/bp_internet/globalbp/globalbp_uk_english/incident_response/STAGING/local_assets/downloads_pdfs/Deepwater_Horizon_Containment_Response.pdf

British Petroleum. (2011). Response timeline. Retrieved from http://www.bp.com/iframe.do?categoryId=9035136&contentId=7065156

Curlee, W., & Gordon, R. (2010). Complexity theory and project management. Hoboken, NJ: Wiley.

Deloitte. (2010). Buying into change: Recovering, reacting, redirecting in oil and gas. Houston, TX: Deloitte Center for Energy Solutions.

Duarte, D., & Snyder, N. (2006). Mastering virtual teams: Strategies, tools, and techniques that succeed (3rd ed.). San Francisco, CA: Jossey-Bass.

Measuring the Chinese fleet. (n.d.). In StrategyWorld.com. Retrieved from http://www.strategypage.com/htmw/htsurf/articles/20100121.aspx

McKinnie, R. (2007). The application of complexity theory to the field of project management. (UMI No. 3283983)

National Commission on the BP deepwater horizon oil spill and offshore drilling (2011, January). Deep water: The gulf disaster and the future of offshore drilling. - The report to the President. Retrieved from https://s3.amazonaws.com/pdf_final/DEEPWATER_ReporttothePresident_FINAL.pdf

Pritchard, C. (2005). Risk management: Concept and guidance (3rd ed.). Arlington, VA.: ESI International.

Project Management Institute. (2008). A guide to the project management body of knowledge (PMBOK® Guide) (4th ed.). Newtown Square, PA: Project Management Institute.

Royal Australian Navy. (n.d.) In Wikipedia. Retrieved from http://en.wikipedia.org/wiki/Royal_Australian_Navy

Weaver, P. (2007, August). A simple view of complexity in project management. Paper presented at PMOZ 2007, Gold Coast, Queensland, Australia.

© 2010, Wanda Curlee and Robert Gordon
Originally published as a part of 2011 PMI Global Congress Proceedings – Dublin, Ireland



Related Content

  • PMI Case Study

    Saudi Aramco member content open

    This in-depth case study outlines a project to increase productivity with Saudi Arabian public petroleum and natural gas company, Saudi Aramco.

  • PM Network

    High-Wire Act member content open

    By Parsi, Novid Seven years ago, government leaders in Alberta, Canada vowed to take a major step toward addressing the energy needs of the growing province. The resulting transmission line, the longest of its kind…

  • PM Network

    Proyecto de alta tensión member content open

    By Parsi, Novid Hace siete años, los líderes gubernamentales de Alberta, Canadá, se comprometieron a dar un paso importante para abordar las necesidades energéticas de la provincia en crecimiento. La línea de…

  • PMI White Paper

    The Impact of the COVID-19 Crisis on Project Business member content open

    By Project Management Institute | Project Business Foundation This is a report on the results of a survey jointly conducted by PMI and the Project Business Foundation. The intention was to replace observations and opinions with reliable data. The mission was…

  • PM Network

    Virtual Count member content open

    By Parsi, Novid The United States government is about to embark on one of its highest-stakes IT projects ever—the nation's first digital census. Households will complete online surveys, and field workers will…