|
|
|
|
 |
 |
February 2005 Organizing for Performance: How BP Did It
by John Roberts Based on “Organizing for Performance,” a chapter in the book The Modern Firm, by John Roberts, 2004. The Economist reviewer of business books called it the best business book of 2004. Roberts is the John H. Scully Professor of Economics, Strategic Management, and International Business and codirector of the Center for Global Business and the Economy at the Business School. This article is published by permission of Oxford University Press. In the last decade, many firms have learned how to organize themselves to deliver exceptional performance. One of the most successful has been BP, the energy giant formerly known as British Petroleum. Examining BP’s story gives insights into one organizational strategy that can achieve outstanding results. In the late 1980s, BP was a politicized, top-heavy bureaucracy managed through a cumbersome matrix structure. The company was spread across numerous distinct lines of business, the result of its not having yet completely undone the conglomerate diversification in which it had indulged in the 1970s. Financial proposals required 15 signatures before they could be accepted; head-office staff filled a 32-story building; and meetings of 86 committees absorbed top executives’ days. Performance was declining, the company was heavily indebted, and by 1992, the company faced a financial crisis that almost resulted in bankruptcy. In recent years, in contrast, BP has reported some of the highest profits ever recorded by any corporation—$4 billion in the third quarter of 2004 alone. While high prices for crude oil contributed to these results, the company’s cost-efficient operations, especially in its exploration and production business, were key to the strong performance it has generated over the last decade. Indeed, BP is widely recognized in the energy industry for its effectiveness in finding hydrocarbon deposits and its efficiency in bringing then to market. This operational excellence was actually maintained and improved in circumstances that could have resulted in major disruptions in the workings of the firm: Between 1998 and 2000, the company absorbed two other major integrated oil companies, Amoco and Arco, that together were almost as large as BP itself, as well as a number of other companies. These acquisitions brought scale and scope, but it was strategic and organizational changes implemented over the preceding decade under three successive CEOs that formed the basis for BP’s remarkable turnaround. The changes that eventually transformed BP began with divesting unrelated lines of business. The company then focused on three basic businesses: “upstream” oil and gas exploration and production, “downstream” petroleum refining and marketing, and petrochemicals. These business lines were obviously related, as upstream’s crude oil is the basic input for the other two streams, but the three were actually run quite independently, because well functioning markets for crude meant there were no advantages to internal dealings. With the corporate strategy made clear, the organizational changes that would generate improved performance began under Robert Horton, who became CEO in 1989. His “Project 1990” sought to improve the speed and effectiveness of managerial decision-making. He transferred authority for many decisions from the corporate center to the business streams. In the process, layers of management were eliminated, and headquarters employment decreased by over 80 percent. Employees were encouraged to take responsibility and exercise initiative. Values of caring, trust, openness, teamwork, and cooperation were espoused. At the same time, as economic difficulties mounted, capital budgets were slashed, and employment was cut deeply. Horton’s abrasive personal style and the dissonance between the proclaimed values and the reality of job cuts alienated employees, while economic performance continued to deteriorate in the context of the general economic slowdown of the early 1990s. In 1992 Horton was replaced by David Simon, who, although more popular among employees, continued Horton’s aggressive cost cutting. Employment fell from more than 97,000 in 1992 to just over 50,000 in 1995, and the company moved from a loss of $811 million in 1992 to a profit of $2.4 billion two years later, while its debt was reduced by $4 billion. The biggest changes during this period occurred in the core upstream business, BP Exploration (BPX). There, John Browne (Sloan ’81), who would succeed Simon as BP Group CEO in 1995, undertook a fundamental organizational redesign. The model, which BPX called an “asset federation,” was later applied across the company as a whole. It led to fundamental behavioral changes and impressive results. Browne began by refocusing the upstream strategy on finding and exploiting really large hydrocarbon deposits where the technical difficulties and attendant risks meant that BP’s expertise and size gave it a relative advantage over “petropreneurs”—smaller firms who were often more successful than the large integrated oil companies. This strategy ran against the conventional wisdom of the time, which held that there were no more big finds to be had. It meant that success would depend on BPX’s ability to discover such fields and to produce the resulting crude at low cost. The next move was organizational. It was designed to induce the entrepreneurial initiative that would be needed under the new strategy. BPX had been divided into regional operating companies (ROCs), which had staffs of technical and business people overseeing actual operations. The heads of the ROCs and functions joined Browne in a global management group that ran the business stream. Managers of the actual fields had very limited discretion and control over the resources used in their units. Browne began by pushing performance evaluation discussions down from the ROC level to individual fields. This led to a conscious experiment in organizational design, with the managers of a number of fields being given authority to decide how to run their operations and how to meet performance targets that they negotiated directly with top management of BPX. When this change resulted in increased outputs and reduced costs, the model was applied throughout BPX, beginning in the crisis year of 1992. Exploration and production operations were divided into some 40 business units, called assets, each of which consisted of a major oil or gas field or group of co-located fields. Each was headed by an asset manager, later called a business unit leader. The ROCs were eliminated, with senior management of the stream pared down to Browne and two others who formed the BPX Executive Committee (EXCO). Technical and functional staffs were largely dispersed to the assets. The asset managers signed individual performance contracts with Browne, agreeing to deliver specified levels of performance in terms of production volumes, costs, and capital expenditure. Within the limits of general corporate policy, they were then empowered to figure out how to achieve their promised performance. They could decide on outsourcing and choose suppliers, do their own hiring, and determine where and how to drill. The performance was not aggregated below the level of the stream itself, and therefore, the performance of individual assets was fully transparent to EXCO. It tracked performance closely, investing in improved measurement and information systems and holding rigorous quarterly performance reviews. Through conversations in these meetings, Browne coached the asset managers, helping them develop their managerial skills and inculcating the values and norms he sought to spread through BPX. Promises made in the performance contracts at the asset level became the basis for performance contracts for everyone within the asset with the asset manager. All employees’ compensation was tied to their asset’s performance and to the overall performance of the stream, which increased the variability of pay and the intensity of incentives significantly. The asset managers found the new system liberating, and they responded entrepreneurially to the increased incentives, authority, and accountability with increased enthusiasm, imagination, and effort, just as had been intended. A problem soon arose, however. The leanness at the top that protected the asset managers from headquarters’ interference also meant they could not rely on headquarters to advise and support them when technical or commercial problems arose. To respond to this need, the assets were aligned into four peer groups that, after some experimentation, were defined on the basis of the life stage of the assets. The key point was that assets within a group, although geographically dispersed, were likely to face similar problems. The groups met frequently, without any EXCO involvement. A system of peer assists was then established under which an asset facing a problem could call on other assets in its peer group to send people to help solve the problem. Numerous other “federal groups” linked people with common interests and challenges across different assets. Strikingly, there were no explicit rewards for this cooperative activity, although it was crucial to sharing knowledge and driving down costs. Instead, the recognized mutual interdependence and the personal relations among the asset managers were the motivation. The peer groups were also given another role early on, called peer challenge. Under it, peer group members were expected to challenge one another on the targets that they negotiated individually with the executive committee. This allowed collective expert knowledge to be brought to bear in establishing targets—knowledge that was not available to EXCO. Later, the peer groups each took collective responsibility for meeting the performance targets of the member assets and for allocating capital among them. At the same time, increasing reliance was put on outsourcing. This extended to activities previously seen as critical, including the generation of seismic data on potential new fields: Only the interpretation of the data was kept in-house. Strikingly, the logic of the performance contracts was sometimes extended to outside suppliers, whose payments were made a function of their performance. A triumph for this approach was the Andrew field in the North Sea, which had previously been believed to be too expensive to bring into production. By sharing cost savings with its contractors, BP was able to develop the field at a fraction of the original cost estimates and in much less time than had been believed necessary. This organizational model led to remarkable successes at BPX. New fields were found and developed, the cost of developing fields was reduced substantially and kept being squeezed, and the productive life of assets was extended. After Browne became CEO of British Petroleum in 1995, this model was applied across the whole company. The appropriate definition of assets was less obvious in the other streams than at BPX, and establishing the right performance measures also presented challenges. Still, the system of discrete business units, peer groups and peer assists, small executive committees for each stream, performance contracts and pay, improved measurement and information systems, and peer challenge was instituted. These changes in the organizational architecture and routines eventually led to fundamental cultural changes. BP’s people developed a deep, intrinsic dedication to delivering ever-improving performance. Strong norms emerged of mutual trust, of admitting early when one faced difficulties (“no surprises”) and seeking assistance, of responding positively to requests for help, of keeping promises about performance. These norms had powerful effects, supplementing the changes in architecture and process and generating remarkable initiative within the assets and great cooperation across them. This model has been adjusted since, but its basic logic persisted through the 1990s and proved especially valuable when BP acquired Amoco and Arco, because the new assets were quickly integrated with the support of the peer groups. In the wake of BP’s success, many firms have sought explicitly to emulate its organizational design, while others adopted similar organizational designs on their own. There is, of course, no one best way to organize that all firms should adopt. The best organization design depends on the strategy being pursued, the market and non-market context, and the administrative heritage of the organization. Nonetheless, real gains in performance can often be achieved by adopting designs that adhere to the basic logic underlying BP’s disaggregated model. These include:
Implementing these—and getting them right—can, as at BP, lead to huge improvements in performance. |
|
Building Cooperation and InitiativeAll organizations need two distinct sorts of behavior from their key members. One is the imaginative, diligent, principled pursuit of the individuals’ own responsibilities, for example, reducing costs and increasing volumes in a BPX asset. The other is the generous, energetic, effective support of others’ objectives and corporate interests, such as sharing knowledge across assets or allocating capital effectively. We call these “initiative” and “cooperation.” The first is obviously crucial to organizational performance; the latter is too, if there are any real interdependencies across units—and if there are not such interdependencies, then there is no reason for the units to be in the same company! Getting more of both is the key to increasing performance, and getting the appropriate balance is crucial to optimizing the organization. The difficulty is that measures that induce more of one tend to result in less of the other. Managers need to think about the levels and mix of cooperation and initiative their organizational designs are producing and whether these are appropriate. If not, they should consider what changes might induce the appropriate behaviors. At BPX, the delayering, creation of assets, empowerment, performance contracts, performance pay, and increased focus through outsourcing all increased initiative in a complementary fashion, with each design feature making the others more effective. However, they probably reduced the willingness to cooperate. After all, the managers had only so much time available, and working harder on their own responsibilities left less time for other things. Thus the peer groups and peer assists were designed to increase cooperation, although they may have eaten into the amount of initiative being provided. On net, however, the new design led to more of both cooperation and initiative.
|
|
|
 |
|
|
|
|
