posted 1 May 2000 in Volume 3 Issue 8
The origin of species: The evolution
of networks at BP Amoco
At BP Amoco, the structure of networks has evolved into an invaluable mechanism for sharing knowledge and best practice. Nick Milton and Walt Palen present a history of networking at the company, and describe the lessons they have learned over the past ten years.
'A networked organisation requires a culture which encourages people to share knowledge readily, regardless of organisational boundaries.' BP Corporate Vision and Value Statement
Networking has been an established working practice in the BP Amoco group (BPA) for nearly ten years. Networks were originally introduced as an organisational change mechanism, but as the company moved to a federal model they became seen as a vital mechanism for preserving the competency of the firm. More recently, the introduction of knowledge management concepts and understanding has allowed networks to become more consciously effective as a knowledge sharing mechanism. Many types of network have become established, varying from formal networks with high-level sponsorship, 'contracted' to deliver products to the business, to informal Communities of Practice (CoPs) dedicated to sharing knowledge around the globe. This paper outlines some of the experiences and early lessons from the use of networks in BP Amoco. A follow-up paper is planned for the autumn, to discuss how the process of networking has evolved since the merger.
The history of networking
BP Amoco is a large global energy company; the third-largest oil and gas company in the world. Its major business functions are exploring for and producing oil and gas, oil refining, chemical refining, marketing and transport of fuels, and solar energy. The company was formed at the end of 1998 from a merger between British Petroleum (BP) and Amoco. It consists of 126 business units in over 100 countries, and employs over 80,000 people.
In the 1980s, BP was a very functional organisation. Much of the specialist work of the organisation was done in functional teams, many of which would be located in centres of expertise. If we take the geological function as an example, the company had a chief geologist located in London, with a staff of specialists (for example, the stratigraphic laboratory) who would provide expert services to the business. Major regional centres might also have their own small specialist teams, who would support geologists working on business projects. The fact that the specialists were co-located meant that they could very easily build up a high level of expertise, which could then be made available to the rest of the company. However, because they were separated geographically from the business teams meant that this expertise was rarely made fully available.
BP recognised the deficiencies of the functional model, and in 1991 began a progressive programme of organisational change. The exploration department started the process, by moving to a federation of business units in their UK operations centre in Aberdeen. This organisational change was accompanied by culture change, and the four components of what was termed OPEN behaviour (openness, performance focus, empowerment and networking) were chosen to set the tone of the new culture. Each of these behavioural components was critical to working in a federal model, and networking in particular was seen as vital for maintaining corporate competence as the organisation decentralised. The specialist teams were disbanded, the company was downsized and the specialists were assigned to business units where their skills could be directly addressed to business problems. This had the massive advantage that the source of expertise was applied where needed, but of course carried the risk that the specialists, no longer co-located, would become isolated, and start to lose their technical edge and breadth by working in singular areas.
To help counteract this tendency, the company set up a number of technical, functional, and commercial networks. These tended to be formal networks, with defined membership and set objectives, and operating through face-to-face meetings (monthly, quarterly or annual). Technical networks organised along common topics, for example that of handling produced water from oil wells. Functional networks included issues specific to the onshore and offshore operations or drilling activity. The commercial networks formed around the need to manage BP's asset based structure via the use of peer groups (roughly organised by asset maturity).
As the company moved progressively to the federal model, the decision was made to implement a common IT infrastructure, known as the Common Operating Environment (COE) based on the 'Wintel' platform. This set of standard software and technology enabled employees across the group to be able to share files, exchange documents and communicate without barriers. It was becoming obvious that the combination of COE and networks gave the BP group the necessary infrastructure and culture to begin systematically to manage its operational knowledge. A group-wide knowledge management initiative was started in 1997, in order to provide standard tools and processes for managing knowledge across the business units. The knowledge management team introduced many of the concepts of communities of practice, and helped to develop BP's understanding of the networks as a knowledge-sharing mechanism - one part of a six-component holistic model for systematic knowledge management now known as KEP. As a result, a large number of informal networks began to be recognised, defined or started up. In some cases these networks already existed as loose affiliations of practitioners, and the CoP concept helped to legitimise them, and to give them direction and impetus.
In late 1998, BP was facing the merger with Amoco. Both organisations recognised the value of networking, and it was obviously crucial that the network should be merged as soon as possible to allow free flow of knowledge across the new organisation. It was also an opportune time to look at the different types of network, and to start the new organisation with a single network model.
The promotion of networking behaviour, combined with the freedom of communication that COE allowed, had resulted in the formation of a wide range of networks. These ranged from very formal networks of business managers, close to being virtual teams, right the way through to very informal networks of people who had worked together in the past and wanted to stay in touch, rather like old-boy networks. There were small networks of a few people, and huge networks with hundreds of members. There were networks that met annually, and networks that communicated daily. However, looking at the range of networks in the latter half of 1998, it became clear that there were two main types.
There were the formal networks, which were sponsored and resourced by the business and which often held performance contracts or expectations, committing to deliver value to the business by working collectively. These networks had defined membership, and a formal programme of meetings, with objectives, deliverables and so on. Examples of this type of network include:
The second main type of network was the more informal sort, close to the established concepts of communities of practice (Wenger, 1998). These were sometimes given resources by the business, often to the extent of funding a network co-ordinator, but did not collectively contract to deliver value to the organisation. There was often no defined membership, and often no fixed programme of meetings, the network preferring to meet continuously but virtually. Examples of this type of network include;
- The green operations network; a global network of people concerned to
minimise the environmental impact of BP's operations.
- The seismic survey network, where people shared knowledge on acquiring
land and marine seismic surveys.
- The grease network; staff involved in the manufacture and marketing of lubricants.
There are now 50 or 60 formal networks recognised in the BP Amoco group, with probably five times as many informal networks. These two types of networks - the formal and informal - work in very different ways. This can best be understood by looking at some case histories.
The produced water network (informal network)
The produced water network (PWN) is the oldest and best-publicised network in BPA. It originated in the North Sea operations area, which long ago began producing more water than oil from its older fields. Participation was quickly adopted by the Alaskan assets that shared the same common issues; namely, how do you best minimise and treat large amounts of water production from facilities that in the main were designed and operated to handle large amounts of oil? The PWN is aligned with some of the remaining technical service programs in BPA's research centre in Sunbury, and has a formal network co-ordinator. Its current membership is over 400 staff, the majority with working positions in operating assets. Its main practice is to circulate, via a discussion folder, any hot topic issues amongst the membership. Responses on questions are rarely vetted, but can be challenged if needed by participants. It also holds a central repository of knowledge concerning produced water on an internal intranet site.
The groups hold regional meetings at the main sites to reaffirm participation as well as to look at technology dominant issues that are hard to convey in email messages. It has been responsible for rapidly moving the offshore assets to common better practices in handling large amounts of water (for example, by introducing hydrocyclones), and now has affiliated networks that look purely at separation technology and practice as well as plant optimisation. The success of this network can be found in both its grass roots organisation and its blue-collar support, as it was in practice before networks were vogue in BPA. It has also had a fairly steady membership and central support as well as a good management advocate.
The Operations Network (formal network)
The Operations Network (Ops Net) was formed in 1994 by the then functional head of HSE and operations, Bill Sears. With the company growing again, management realised that a senior network would be needed to keep pace with external expectations on HSE, Production, and Lifting costs. The group was charted to meet and to provide peer reviews bi-annually at the different operating assets locations, and also met monthly by videoconference. The larger asset groupings had their own local chapters of the Ops Net, which looked after the contributions to the larger company as well as managing local issues that were not common. The representatives of these meetings were usually the operations managers, second only to the asset manager/business unit leader in the asset hierarchy, and included senior discipline representatives from the mostly UK based technical service groups.
In 1997, the Ops Net was asked to put together a list of objectives that they would be held to in delivery, and this was mirrored on the local level as well. In 1998, these objectives became a more formal charter for the group. This in itself caused controversy since in the BP federal model, only assets (or their peer groupings) deliver hard benefits, like production, and it was impossible to get closure on issues of prioritising what the network would actually deliver. Now (2000) the network has formed into many sub-groupings, which are able to look at specific issues with greater focus, like production efficiency, and left the main group to concentrate on other issues, such as promotions and succession planning.
The 3D modelling network (informal network)
3D-reservoir modelling is a technique used in BPX to improve the understanding of oil and gas fields, by building detailed computer models of the underground reservoir layers. The conviction is that a better reservoir description leads to better reservoir management and so to better reservoir performance (more oil, faster oil, less water and so on). These computer models are extremely detailed and complex, and may contain many million cells. 3D modelling is a fairly immature science. It only started to move out of the research centre into the business units in the early 1990s. Recently, the software vendors have been producing software for the task, which anyone can use. This has broadened the use of the technique, but exposed the immaturity of the science, and resulted in a growing number of practitioners with a need to learn from each other. A conference in April 1996 highlighted the need for a network of reservoir modellers, and Ray King (an experienced geophysicist from Aberdeen) volunteered to organise the network. He based the network processes around an email discussion forum, initially on a list server, but moving to MS Exchange public folders after COE was introduced. Ray recalls:
'Once we had it rolling in October 1996, it took off, and there was plenty of e-mail traffic. People got the hang of being able to reach over 40 technical professionals reliably, and get messages back. They can participate in the discussion, or can view the messages without committing to being on the list. The visible network, that is, those who are on the mailing list, now contains over 100 practitioners, some well-versed in 3D-reservoir modelling, others keen to learn. Many are practitioners in remote offices; places like Caracas or Kuwait.'
Currently, the 3Dmod head count stands at 98, in locations such as London, Aberdeen, Caracas, Houston, Anchorage, Kuwait and Bogota. There is also an invisible network, people who access the public folder and participate passively: They may decide to join in visibly, or not. If they do, they subscribe voluntarily by sending a note to Ray, and he puts them on the mailing list. For a year or two, the knowledge generated by the 3D Modelling network resided, unsorted, in the many hundred of messages in the public folder. Ray and the network realised that this was not the ideal solution; knowledge was becoming hard to find, and the same questions were being raised time and again. The 3D modellers network have now set up an intranet site, where they store and share distilled knowledge, frequently asked questions, and best practice. The network receives no funding from the central organisation, but involves very little expense. Ray King spends about 5% of his time organising the network, which is largely self-sustaining. The members contribute their own time and expertise to the discussions, and in return gain timesaving tips and solutions. The network holds no performance contract, and delivers value by increasing the efficiency of the members, so they can build complex 3D models in rapid time.
The tools of a network
There are several tools and enablers that are common to the BP networks, and which are essential to their success. The most important enabler is to have a network facilitator, somebody to co-ordinate the activity of the network. Experience in BP Amoco suggests that the facilitator needs to be very much an 'insider' 'someone who understands the issues, knows the people', 'speaks the language', and has the trust and respect of the network members. For example, the role of Ray King as moderator of the 3Dmod group has been crucial; Ray administers the mailing list and public folder, monitors traffic, and sparks debates if things go quiet. Similarly, Frank Sweeney supports the Produced Water Network with both physical knowledge and funds, and Andrew Goodwin supports and leads the Operations Network activity in Production Efficiency.
The network also needs a means of communication. A co-located network can always communicate face-to-face, and local networks such as the Houston geologists often arrange regular meetings and talks and 'brown-bag lunches' where they can get together and exchange knowledge. However, the vast majority of the networks in BP Amoco are not co-located, but contain members from many different countries. The most effective means of communication has proven to be email, using public folders in MS Exchange to build-up discussion groups and to share documents. Email has the advantage of being an established work practice, and is part of everybody's communication habit. Some networks experimented with web-hosted discussion forums, but these quickly died, as people got out of the habit of visiting the website to join the discussion.
The discussion forum for the Produced Water network is displayed within a personal email window. The discussion folder is in the left-hand panel, and appears in bold whenever it contains unread messages. Opening the discussion folder shows a series of conversation threads, where an initial question is followed by answers and discussion from other network members. The number of replies to each question varies from one or two up to about 30 in this particular forum, depending on the degree of knowledge and the degree of interest of the network. Generally each question receives a reply within a couple of days, and if not, the facilitator will generally try and intervene.
The network needs some common storage facility for community knowledge. A website most effectively provides this, as everyone in the BP Amoco group has intranet access. Maintaining a web-site is a time-consuming task, and it is generally the networks with centrally funded resource that are able to maintain large repositories of knowledge. Networks such as Green Operations, Produced Water and the Grease network host well-populated websites where network members can find resources to help them do their job.
Occasional face-to-face meetings are needed to keep the network fresh, and to maintain relationships between the members. The knowledge management community held a very effective high-energy kick-off meeting in Milan during late 1997, but was unable to meet again for two years, and it was obvious that the original impetus had been lost. Annual meetings of the core membership seem to be necessary to maintain drive, even in informal communities.
An additional enabler for the networks in BP Amoco is the corporate Yellow Pages system known as Connect (Collison, 1999). This interactive web-based system allows each individual in the company to write their own home page; complete with an inventory of their skills and knowledge, and a list of the networks they belong to. Connect can be searched for skills, so a provisional membership list for a new skills-based network can be compiled very quickly. Connect also acts as a directory for the existing networks. A search for the 3D modellers network, for example, would pull up a page showing the network description, the co-ordinator, the number of members, and a link to the network website, and then a table of all members with links to their individual connect pages. This table could be sorted by name, location or job title, and the thumbnail photographs (where available) of the network members could also be displayed.
The way they work
The formal networks tend to mirror the asset business practices. Meetings are scheduled ahead of time, delegates chosen/agreed, agendas created and updated and speakers prepared. In line with the conduct of most operations groups, there is usually time allocated in the meetings for all delegates to share either opinion or knowledge on the focus topic for the meeting, for example, compression or separation. Each asset representative is also asked to bring one key issue from their own asset, which they can share as better practice or to seek assistance from the group. Then a peer assist is held, on behalf of the asset that is hosting the meeting, usually attended by the senior members of the operations network. The topic will be a general operations review for internal compliance or on the asset's specific operational issue causing difficulty e.g. separation train performance at high altitudes. This meeting is scheduled to be concluded before the formal agenda discussions begin. That way immediate lessons and observations from the peers are shared with both the asset and the larger operations community attending the meeting, who then have opportunity to offer assistance immediately, and on the spot, if needed. Routinely, tasks are then given, targets are measured/updated; minutes issued and people resume their daily lives. Often the technical specialists will update the host websites so that the attendees at large and local chapters can share meeting findings and outcomes.
The informal networks, or communities of practice, operate through rather different principles. The primary raison d'Ítre for these communities is to act as mechanisms for the exchange of knowledge, in order to increase the effectiveness of the members. The value proposition for the community is at the level of the individual members. If the value they derive from the community is greater than the investment they need to make (primarily an investment of time and effort) then they will stay in the community. If it isn't, they will leave. So freedom of membership is an important principle in the informal networks. Members have to be free to join or leave depending on the value. They effectively vote with their feet.
Knowledge can only be exchanged in an atmosphere of trust. The community members need to feel confident that they can offer their advice and ideas without being shot down in flames. The people who ask the questions and receive the advice, need to be able to trust the advice and recommendations they receive. Building this relationship of trust and mutual respect is one of the more important roles of the network facilitator. The face-to-face meetings are instrumental in relationship building, but relationships can be built vie email as well (witness the number of people who get engaged or married via the Internet), with a light touch from the facilitator. Ray King explains:
'I encourage people to overcome their shyness in airing things that may not be earth-shattering, but still may be of interest to others, using gentle persuasion, but no kind of pressure. I leave it up to what comes naturally.'
There is no defined hierarchy among the BP Amoco communities of practice. The email discussion groups are open to all, and everyone is equally free to raise questions or to offer advice. Nobody is designated 'expert', though some of the networks define people as accountable for maintaining the community knowledge base in a specific area of expertise (the BP oil refining term for these people is 'knowledge area owners'). In practice, however, there is a subset of people in each community who answer most of the questions. These are either the extroverts, or the people who really do have deep practical experience.
The most prolific contributor to the knowledge management discussion forum was the network facilitator, who played an active role in steering and summarising discussions. Another 99 individuals contributed to the discussions. This was more than half the total number of community members, so the number of 'lurkers', or non-participants was about 70. Half the people who contributed, only contributed a single message, which was often a request for advice. The '80:20' rule seems to apply to this discussion group, as to many others, in that roughly 80% of the discussion involves 20% of the members. However, the caveat is that much discussion takes place directly between the members, and not through the public folder.
BP Amoco invested heavily in nurturing and creating networks and COPs, primarily to respond to the business challenge of moving to a distributed asset model and forced downsizing. Two generic types of networks can be identified; the formal or management-driven networks that are focused on the current business environment issues and which hold collective performance contracts, and the informal staff-led networks that focus on discipline issues and which deliver value through enhancing the knowledge base of the individual members. The most successful networks have strong central support and growing membership, the less successful networks may constantly be changing formal members as management tunes the topic and objectives.
So how does the company decide which networks should be formal, and which should be informal? Because the formal networks have a defined customer, with whom they contract performance, they are generally defined by the businesses. A strategic knowledge need that faces a business unit or (more commonly) a peer group of business units, can be addressed by setting up and resourcing a formal network. Informal networks either spring up on their own, in a response to the needs of practitioners in the group (as in the 3D modelling example), or else nucleate around a team (such as the knowledge management community), an expert (PWN), or around a formal network. It is quite common for an informal network of interested people to coalesce around a formal network to disseminate the knowledge that the formal network develops.
Collison, 'Connecting the new organisation - how BP Amoco encourages post-merger collaboration', KM Review (Issue 7, March/April 1999)
Nick Milton is the managing director of Knowledge Transform(SM) International. He can be contacted at:firstname.lastname@example.org
Walt Palen is VP and director of KTI for Commercial and Technology. He can be contacted at:email@example.com