Bo Dahlbom, Göteborg University

Ole Hanseth, Oslo University

Jan Ljungberg, Viktoria Institute

SKF is world leader in the production of ball and roller bearings, having 15-20% of the world market. Ever since it was established in 1907 it has been a global, production and distribution focused company, selling the very same type of products all over the world. Of course SKF has changed throughout this period. However, what is most striking with the company—in particular compared with the picture of modern global organizations painted by current management literature—is its stability. SKF is changing, but in small steps.

Just as the overall company, the information technology infrastructure has evolved slowly and in relative harmony with the organizational changes. This chapter will outline the successful co-evolution of SKF and its infrastructure. We shall see how, in the last thirty years, SKF has relied on a strategy for building infrastructures focusing on standardization and hence stability, in effect having an institutional perspective on its organization, focusing on inertia and the cultivation of entrenched infrastructures (Hanseth 1996). This is in obvious conflict with modern management literature focus on dynamics and flexibility and on radical and rapid organizational change, designing infrastructures from scratch.

Even when SKF seems to be following modern infrastructure strategies, aligning business and infrastructure strategy, following a top down, centrally controlled, approach, it does so by developing slowly, without radical breaks with its past. But how long will such a conservative approach be possible? Can SKF meet an increasingly global competition by simply making its production and distribution organization more efficient? Or is there a need for more radical changes at last?

SKF was founded in 1907 by a young engineer, Sven Winquist, at Göteborg, Sweden, but immediately became a multi-national company. Its main product is ball and roller bearings. In addition, SKF produces steel, and seals. The seals are to a large extent produced to match the bearings and the steel is raw material for the bearings. The bearings' part of the company (which is the subject of this chapter) is organized in 6 divisions: Automotive, Electrical, Industrial, Asia and Pacific, Specialty Products, and Emerging Businesses. The current strategy focus is on growth in the growing markets in Asia and Central and Eastern Europe.

SKF’s sales in 1998 reached SEK 38 billion (in 1996 SEK 34 billion), and the number of employees 45 000. The company is operating in more than 130 countries. Production is taking place at more than 82 plants, in 54 locations in 20 countries. Customers are reached through 10 000 distributors and approximately 100 SKF sales offices. Due to investments of SEK 3 billion in a global restructuring program, there was a negative balance of SEK 2 billion in 1998 (in 1996 the profit margin was SEK 2.4 billion).

SKF produces (and sells) about 22 000 different bearings. The smallest one is about 2 mm in diameter (used in airplane gyroscopes). The largest one is about 7.2 meter in diameter, and its weight is about 45 tons (used in tunnel drilling machines). In its first year SKF produced 2 200 bearings. Currently it produces 227 000—per hour!

SKF considers its strengths to be technical superiority, quality, and a world wide distribution system. A well qualified work force of application engineers is ready to solve customer problems, give advice, and provide solutions. The high quality is only possible because of excellent engineering competence. Thus, much of the production equipment is manufactured by the company.

SKF began in 1907 with 15 employees and a yearly production of 2 200 bearings. In its first three years, it was established in France, Germany, England, Finland, Australia, and the US. Then SKF moved into Norway, Belgium, Holland, Russia, South Africa, China and Austria. In 1918, SKF had 12 factories around the world, 12 000 employees, and sales in 100 countries. Accordingly we might say that SKF has always been a global company. However, what it means to be global has changed—and is rapidly changing now in the late 1990s.

One way of being a "global" company—in fact the most accepted one—is global presence, primarily by selling the products globally. Strategies for organizing and operating SKF as a global organization has changed as transport and communication technology have changed. It has also changed as SKF’s customers have become more global.

SKF has continuously globalized through expansion. Although the company always has been global, its activities have remained based in Göteborg. The closer to Göteborg, the larger the market share. For instance, in 1996, 72% of production and 55% of sales were in Western Europe. On the other hand, SKF activities in this area are not growing. The growth is located to the new markets in Central and Eastern Europe and Asia.

In Central and Eastern Europe, SKF is to a large extent expanding by buying factories earlier run by the governments. In 1998, for instance, SKF was taking over a factory in Ukraine. SKF has had production units in operation in India since the 1960s, in Singapore since the 70s, in Japan since the 80s, in Taiwan and Malaysia since the early 90s. In China, a new plant began production in Shanghai in early 1997, another one in Nankou (together with the Ministry for Railroads in China).

Currently, expansion into new markets primarily takes place through joint ventures. For instance, SKF has bought a 20% share in the leading bearings manufacturer in China, Wafangdian Bearings Company. SKF and this company are together setting up new joint venture production companies. SKF will be majority owner and responsible for leadership in these companies. Similarly, SKF is establishing a joint venture company together with ANZAG for the production of seals. A new factory started up in early 1997 in Korea. This is also a joint venture where SKF is majority owner and responsible for leadership.

In the United States, SKF is strengthening its position by alliances. Thus, a new distribution concept has been introduced together with Rockwell Automation. Since the companies are serving their markets through the same distributors all over the US, SKF and Rockwell has formed a transport alliance, meaning that their products are distributed together from one warehouse in Crossville, Tennessee. There are possibilities for including new companies in the logistics system.

When establishing joint ventures together with local companies, investments in factories, machines equipment, etc. are not the most important aspect. What counts are rather investments in quality education, competence, information and training. Competence development is becoming increasingly more important.

In its early period, the "global" SKF was organized as independent, local (national or regional) organizations. These organizations were self-contained in the sense that they included all functions: production, distribution, marketing and sales, having their own plants, warehouses and sales offices. Only product development was centralized at corporate level.

With time, improved transport and communication technology enabled coordination as well as better distribution services across longer distances. Eventually this led to a major shift in strategy. In the early 1970s, it was decided to introduce an organizational model in which all production and distribution were integrated. The mother-daughter organization was to be replaced by a divisional organization. The idea was to continue being a full range producer of bearings, but cut production costs by specializing production. In principle each bearing should only be produced at one place, in one plant. That plant should supply all customers, globally.

Of course, the change from independent self-contained national or regional companies into one such globally integrated company could not be made instantly. And in keeping with the SKF ideology to introduce changes slowly and carefully, this change is still going on. The new organizational model was not implemented immediately, but rather served as a long term vision for change. Once the change was begun, it has taken a course of continuous, step by step, integration of activities, by extending the area within which activities were operated by one coherent unit.

Growth in the new markets requires a global production, distribution and sales organization. The customers to be served by SKF are becoming more global. Accordingly it is important for SKF to be able to respond to customer needs, wherever the customers happen to be.

In addition to globalization (through expansion and integration), SKF has of course always worked on making the organization more efficient through increased productivity, less products in stock, and improved customer service through increased reliability (delivery at the time promised) and shorter lead times.

The decision to adopt a new organizational model in the 1970s also had implications for IT strategy. As long as the national (or regional) companies were independent and included all functions, they had developed their own IT strategies and solutions. The new integrated model also required an integrated IT infrastructure.

The first IT related decision to be made was to start the development of GFSS — Global Forecasting and Supply System. Such a system was necessary to increase coordination of production and distribution. This system was to be shared by all units in SKF, so it required a shared underlying infrastructure linking the units together. The decision also said that SKF should as far as possible build their own infrastructure rather than use commercially available infrastructural services. The infrastructure was considered so important for the business that the company could not depend on external providers for the necessary services—except for the most basic telecommunication services.

The infrastructure was implemented based on a global SNA network. In fact, SKF was a pioneer in establishing a proprietary international data communications network. The first set of leased lines were established in 1976, providing rudimentary mainframe-to-mainframe connections (Hagström 1991). In 1977 the decision was taken to build a proprietary data communications network.

A number of communication services were established, such as file transfer, remote job entry and message transmission. In 1979 the SKF Group Telenet (Teleprocessing Network) was created to provide basic telecommunication services. Five computer centers in five countries were included: Sweden, West Germany, Italy, France and the UK. A set of corporate applications and systems were supplied as a part of Telenet. One example was the message transfer system MEST (SKF Electronic Message Transmission). The first backbone network was linking the five main European bearing manufacturing units, by a set of leased lines from Göteborg to Schweinfurt, Turin, Clamart and Luton. In 1985 the large computer office at the head-office of bearings in King of Prussia, Pennsylvania, was tied to the network through a leased satellite link.

By tradition, computerized information systems had been handled locally. With the establishment of Group Telenet, those days were gone (Hagström 1991). Coordinated production and distribution required shared information services. In principle, coordination would be best supported by just one global system. However, centralizing all computing to one unit was impossible. A more distributed model, including some local computing resources, was necessary. While distributing computing power, there would only be one system in use. Such a strategy was defined in terms of "Common Systems" for all units. All the different systems in operation were replaced by a portfolio of systems common all over the corporation. The purpose was to link customer orders to production scheduling by inter-communicating systems along the value chain. The new systems portfolio was named the New Material Flow Concept (NMFC). This Common Systems strategy is still fundamental in SKF’s IT strategy.

The set of integrated Common Systems has been a very important infrastructure for SKF for almost 20 years now. In the vocabulary of strategic alignment (Weill et al. 1996), these systems primarily constitute a dependent infrastructure. However, the infrastructure has been enabling as well, since organizational integration has increased considerably since the first version of Common Systems was launched. Furthermore, the systems have served as a utility reducing cost through integration, shortening of lead times and decreasing stock. We will here briefly describe the most important systems in the Common Systems portfolio, from the customer side to the supplier side:

• SCSS – The Sales Company Service System
• ICSS — The International Customer Service System
• MPSS — The Master Production Scheduling System
• MCSS — The Manufacturing Customer Service System

Figure 1. SKF Common Systems

SCSS — The Sales Company Service System was first installed in Brussels in 1987. It encompasses all administrative functions in a sales company except salaries, i.e. it covers marketing support, sales budget and support, customer order handling and invoicing, warehousing and accounts, and even transport scheduling. The main function is to allocate products to customers. By early 1990 it had been implemented in all countries with main node national networks.

SCSS is installed at sales subsidiaries local computers (AS400), which also function as peripheral nodes in the data communications network. Distributors, agents and local sales offices can place orders, check delivery status and make inquiries. SCSS will first try to fill an order from local bearing services stock. If this is not possible then SCSS automatically continues to stocks belonging to the bearing industries warehouses. The SCSS system cannot answer no, it can only answer when bearings are available. If bearings are not available in warehouses, the specialized manufacturing unit replies when the required bearings will be produced.

ICSS — The International Customer Service System was installed in 1982. It is the key interface between the sales and manufacturing organizations, i.e. it books orders against internal supply not available from the domestic warehouses. The main functions are order entry, delivery time fixing, pricing, order acknowledgment and export documents creation (including allocation, packing, invoicing etc.). By 1983 all manufacturing companies in Europe had their own installations (mainframes with MVS). All organizational units connected to Telenet can use ICSS, by having batch as well as on-line connections for transmission of orders, packing specifications and invoices. For the time being the on-line facilities was pretty advanced.

MPSS — The Master Production Scheduling System is a system for overall planning and production. MPSS was installed 1987-1989 in the main European manufacturing companies. The intention with MPSS was to improve production flexibility without losing any of the efficiency improvements gained by production specialization. MPSS identifies resource requirements and checks them against available manpower and machine-capacity, and then generates a master production plan, attempting to keep the scope for possible intervention as wide as possible. ICSS can intervene in the process about one month before assembly. The MPSS system is both guided by information from ICSS and PFS — the Product Forecasting System, the latter calculating product forecasts and safety stocks.

MCSS — The Manufacturing Customer Service System supports shop floor activities, i.e. the day-to-day running and costing of manufacturing. Implementation of MCSS started in 1986 and the main European bearing production companies had it installed by the beginning of the nineties. MCSS schedules operations, checks on capacity and work-load, monitors work and resource consumption, records changes in inventories etc. MCSS regularly generates reports, which are sent to the divisional head-office. MCSS is running on mainframe (MVS), and the company-wide responsibility for maintenance and support is assigned to the Turin node.

When the Common Systems were installed, significant organizational change took place in the form of global harmonization and integration of activities. However, change in organization and infrastructure continues. We will here look at the two most recent major change efforts, both taking place in the 1990s: (1) the introduction of the production channel concept and (2) the "New European Distribution Structure" project (NEDS).

The concept of "production channels" was developed and implemented over some years around 1990. Up to then, SKF produced bearings according to (as they describe it themselves) a Taylorist model. This meant specialization to get economies of scale. Each plant (or production unit) produced only one piece, for instance balls of one specific size, inner rings of specific size and material quality, etc. Then the specific parts were delivered to the locations where the specific bearings were assembled. This model was considered inappropriate and replaced by the channels model, meaning that complete bearings are produced out of the raw material at each channel.

The production channel concept was introduced in the large size bearings factory in Göteborg in 1990/91. This factory produces bearings with a diameter of more than 50 cm. Producing such bearings is something of a craft since each bearing might be unique or produced in very small series. In general, the smaller the bearings are, the more automated the production process will be.

The manufacturing of a large size roller bearing includes the following steps: cutting on the lathe; ultrasound examination, looking for possible defects in the material (This is sometimes a two step process. If one cannot draw a clear conclusion after the first "automatic" check, a closer manual examination is made); hardening; grinding; polishing; assembling the pieces together into a complete bearing; and packaging. Between each operation, the bearing components are stored in buffer areas in the factory. According to the general model there is one area for components on which the operation is finished and one for incoming components. However, the areas are usually "merged" so that one area containing outgoing components from one operation at the same time is the buffer area for incoming components for the next.

The channel system was introduced in two main phases. In the first, the channels were established, using only existing production equipment. This implied that several channels shared equipment, making production process optimization and control quite difficult. The next phase was to introduce new machines so that every channel had their own machines, and to tune the processes. This increased the production capacity. It also made coordination between channels, and accordingly production planning, significantly easier. After the introduction of the channel model, production time in some factories was reduced by 50%.

The introduction of the channel system of course implied that the production related information systems had to be changed as well. In the large size bearings factory, new production planning systems were installed in the late 1980s. These systems were designed to contain information about every piece in the production, registering everything done with it at every station. This would enable managers and machine operators to get an overview of the jobs to be done at every station–how many pieces in line and what to do with them. More aggregated information could also be produced concerning overall production load and estimated output in the next hours, days, weeks, and so on.

However, the information systems were not put into operation in this extensive way. To do so would require lots of information to be entered by the machine operators, and a PC (terminal) was required at every station. Production managers were convinced that this would cost more than it would pay. Instead a more simple production control system has gradually grown out of a combination of old ways of working and the new systems.

Orders are collected from a computer system and the production managers use the orders as a basis for putting together a six-week production plan, based on the capacity of the production channels, filling in with demand estimates to round off production. The six-week plan is a simple sheet of paper, sometimes made on a word processor, sometimes produced by the computer system (in which case, the production manager in case has made his own little program to produce it). The production plan sheet is distributed to all the operators in the channel and it is the basis for discussion at the weekly general channel meeting when stock is taken of the situation. The operators will mark their sheets as they produce, and they will share this information at the meetings.

The actual activities on the shop floor are controlled using a simple idea about the "material as information carrier." Based on this idea, what the production managers call a "material information system" has gradually evolved, with the following elements:

1. Just looking at the amount of material stored in the various buffer areas, gives the information needed about possible bottlenecks in order to utilize the machines optimally.

2. Information about what should be done with the individual pieces is written directly on the material. This covers: (a) marking a special area (using a piece of chalk) that contains a defect and which should be checked later to see if it disappears as the material is cut down; (b) after the cutting process, inner and outer rings are paired together and given identifiers (post-it notes stuck to the surface); (c) after the grinding process the rings are marked (with ink pen) to indicate how much extra material there is left before the polishing. The specification will determine how much of the material will be removed in the polishing. Each bearing is an individual.

This production control system seems to be very simple and is working well. The interface between the material and the computerized information system is simple (a sheet of paper). It supports all operations by introducing as few additional activities as possible. It is very simple and easy to use by the production workers at the shop floor. And it is very cheap.

It is a heterogeneous system including humans, computers, paper forms and the metal material of the production process itself. In this way, the steel rings are not only the objects to be worked on and "manufactured," they are also acting subjects in the production process, informing the operators about what to do.

The engineer in charge was extremely happy with his system and was adamant in pointing out how much better this "material information system" was than the computer based production control system (MCSS) people wanted him to use instead. But looking a little closer at the situation may well make one wonder. The six-week production plan sheet is, of course, the interface to a system that might very well have been run on a computer. And if the operators had to make their marks on such a computer interface rather than on the sheet of paper, then the information about the production process could have been read off the computer rather than being shared at weekly meetings at which people read from their sheets. It is all very well with the buffer areas as a means for calling bottlenecks to everybody’s attention, but without the sheets it is impossible to know how production is really going.

This example also illustrates the fact that a shared infrastructure based on integrated "Common Systems" is not constituted by one universal system installed in an identical way at every location. The different factories are indeed different. To make a shared system work, every unit has to implement its part of it. In doing that, each unit might choose to do that in a way optimizing local needs, or in a way optimizing "external" needs. The large size bearings factory chose the first alternative. By using paper sheets rather than the common production control system, the large size bearings factory has really taken control of its operation away from the main office. As long as it is only on paper, there is no way the production can be closely monitored from the main office. The material information system really has the same sort of raison d’être. By only marking the material itself, all information about the actual production process is kept on the shop floor, away from the eyes of the main office. This material information system was optimized according to the needs of the local production manager. However, it did not contribute in the best way to a shared infrastructure for the whole of SKF.

There are two sorts of systems compared to each other in this case: (1) a material information system with manual routines advocated by engineers and operators on the shop floor; (2) a computer based, powerful, somewhat old-fashioned, production control system advocated by the main office. In our discussions at SKF only one sort of argument was heard for or against these systems: the argument of practicality. It may seem obvious that the powerful computer based production system is impractical in a craftlike production, but this is exactly the issue.

To the extent that the manufacturing of bearings retains elements of craft, then computer based systems will seem impractical. If we side with the engineer on the shop floor, then we seem to have a good argument in favor of the material, manual system. But if we take the perspective of the main office, with the ambition to further automate production, then we might very well say that the disadvantage of the material, manual system is large enough to warrant a more forceful introduction of the computer based system. And, if that system could be made more like the manual paper sheet system they use today in the factory, then perhaps it would even be more practical. And, we might even save time by not having all those meetings!

The manufacturing of bearings at SKF is undergoing continuous automation. The level of central control is such that there is room for decentralized autonomy in the factories, where the central information systems are not used to their full capacity: "We have common systems, but lack common policies, common procedures." In some cases, like in the production of large size bearings, the craftlike elements of production are still strong. And in most factories there are several generations of production technology working side by side. Thus there is room for further automation and more intense use of information technology.

In the 1970s a major effort was made to coordinate production and distribution in Europe. This activity was a part of the implementation of the new organizational model described above. International warehouses were built and production was specialized for the different factories in Europe. All of Europe became one integrated production and distribution system. In the early 1990s this production and distribution system was further refined, in the so-called NEDS ("New European Distribution Structure") project, 1992-1996.

When the NEDS project started in 1992, the 7 factories in Europe each had their own international distribution warehouse, and all the 18 local sales offices in Europe had their own warehouses. When the project was finished, 4 of the factories had international distribution warehouses and all the local warehouses were shut down while one European Distribution Centre (EDC) had been built in Tongeren in Belgium. Big volumes (few order lines) are now delivered directly from factories and the 4 international warehouses, while small volumes (many order lines) for the aftermarket are delivered from EDC.

The new distribution system required a new information system architecture. Before NEDS, every sales unit had its own complete version of the Common Systems for handling customer orders with EDI connections to the many factory warehouses. With NEDS a more centralized system was introduced, described as a client-server architecture, with clients for order and purchase handling at the sales units and central server systems for delivery at the distribution centers. Order handling and delivery now involves several systems which have to communicate:

1. An order is registered in the sales order system at a sales unit. In this system there is a customer data base with information about customers, pricing, and a product data base.

2. A message is sent from the sales order system to one of the distribution centers (selected by the system depending on customer and product) and the order is booked.

3. The server system returns a message to the sales unit, confirming the order and giving time of delivery.

There are about 20 installations of the sales order system at sales units around Europe, and there are 5 installations of the distribution system (at Tongeren and at the 4 international warehouses). New installations are easy to implement, new sales units and distribution channels can be easily introduced independently of each other. The new system provides improved information about products, delivery times and order status directly to customers, plants, distribution and sales units all over Europe. Statistical reports show number of messages sent, types of messages, sender, receiver and process times.

This project has been very successful. It was well planned and preceded by benchmarking. Other international companies, including SKF’s bearing competitors had already performed similar changes with good results.

NEDS system structure, including Warehouse support system (WASS).

Being one of a handful of quickly growing industrial companies when Sweden was industrialized, and producing a vital component in virtually all machinery, SKF was a major force in the industrial revolution. In many ways it has been an exemplary industrial company. Focusing on production and distribution of its products, with large factories and small offices, SKF was built by engineers and has been run by engineers. To this day it remains a production focused company. The educational level is relatively low, and the company has a typical blue collar feeling. What will happen to a company like this in the information technology revolution?

With the introduction of information systems in the 1970s and 80s, offices and office work began to grow. Many engineers left the factories to begin to collaborate more closely with economists in the offices. In many industrial companies there was even a power shift from engineers to economists. This trend grew stronger with the document management systems of the early 1990s, increasing the amount of document processing in many companies, making them focus even more on office work.

SKF has been slow in following these trends. Its information technology use remains focused on production and distribution—production planning, ordering, logistics, and so on. Its offices remain relatively small. Even if it has introduced document management technology like Lotus Notes, it is not really used. In many ways, SKF is an industrial company that has used information technology to support and strengthen its production orientation. Information systems have been introduced as part of more general organizational trends towards centralization and globalization of production and distribution.

So far this strategy has been both necessary and successful. With large industries (automotive companies, electrical appliance companies, paper mills, and so on) as its major customers, the SKF product line can develop relatively slowly and the profit margin is small, so there is neither need nor money for a large office organization. What is important is to be able to continue to produce high quality bearings more and more efficiently with less and less production disturbances. This means that all infrastructure changes have to be introduced slowly and with the active collaboration of the engineers involved, taking into consideration the old infrastructure and giving room for individual bricolage and preferences.

And in this way, SKF has been steadily developing its organization and infrastructure in the direction of a more and more closely knit global production unit for bearings. Nothing that information technology has brought so far, has really changed this traditional approach. SKF remains an industrial company with the same old ambition: to produce high quality bearings and deliver them reliably and quickly, meeting competition by automating and cutting costs.

This traditional approach has worked well and kept SKF at the top among its competitors. But one may well wonder how long this approach will continue to be successful. Now that information technology is rapidly turning into a global market place, with electronic commerce, one to one marketing, and an increasing focus on sales and services rather than production and products, industrial companies like SKF begin to lose their power. No matter how efficient they become at manufacturing and distributing their products, their profit will remain slim. So what shall a company like SKF do?

The standard introduction to consultancy books and articles in the early 1990s told about how things were changing. In the past, they would say, it was possible for companies to rely on a stable market with faithful customers. But all that is different now, they went on to say, with global competition on a global, deregulated, and open market, empowered customers demanding tailor made products and services of high quality, a rapidly developing information technology, more knowledge intensive work and therefore a more educated work force.

If in the past, the most successful organization was the huge factory, mass producing one and the same product to the same customers year after year, we were now entering a situation in which the only surviving companies would be the ones that were able to quickly adjust to a an increasingly changing market with changing demands for new products and services. The recommended recipe for survival always involved moving away from the industrial factory organization towards a more loosely organized, fragmented enterprise with alliances, external networking, and outsourcing as major ingredients (Peters & Waterman 1983, Drucker 1988).

The machine bureaucracies typical of large industrial companies were well suited for the demands of the industrial age. Companies producing large quantities of bearings for other industrial companies rather than for consumers ought not to be particularly worried by the warnings of 1990s consultants. Thus, it may perhaps be a bit silly to take seriously the warnings of consultants, suggesting a development which will turn bearing producing companies into mobile, networking, knowledge organizations.

SKF is still focusing on production, and this means that managers close to the production in the factories are empowered to influence the actual nature of the company infrastructure. The culture of the company is such that the experience and competence of the engineers directly involved in production shapes the whole company. The way Common Systems have been implemented is a good example of how SKF remains a company where the head office is careful not to enforce its ideas on the engineers and operators on the factory floor. As an indication of the culture of SKF, everybody speaks about production and distribution when they describe their company.

With industrial companies as its major customers, the market is relatively easy to manage. Even if SKF has 25 000 customers, this is really a small number, compared to the global market of a pharmaceutical company or a PC company. And among these 25 000 customers there are very few consumers, if any. Thus when we look at the global challenge to turn into a customer and service oriented, distributed, mobile, sales organization, this does not really affect SKF.

And yet, this may not be the whole truth. Globalization means increasing competition, and when several companies are selling similar products, prices will drop and only one will survive. Surviving means focusing on "adding value" to the customer, i.e. offering her a product satisfying her needs in a way that competitors cannot match, and for which she is willing to pay a good price. In Reich’s (1991) view, those surviving in the global competition are turned into service companies. Service delivery is the opposite of mass production. Key characteristics are customer or client interaction as the service is consumed when it is produced.

SKF today is far form being a service provider. However, there are some indications that it is moving towards the images presented of modern, global organizations. "Customer in focus"—has been a slogan for 10 years. Now it starts to mean something. Walking through the factories, there are all sorts of signs of a growing customer awareness. (Hanging from the ceiling above the production channels, there are signboards with the names of the current customers of each channel. Asking one of the foremen to name the companies on the board, without looking, he got 2 out of 4 right.) These signs may seem innocent, but if they are taken seriously they point in a direction which will break up the integrated production and distribution system of SKF, strengthening direct lines of communication between customers and individual factories, and thus threaten the traditional production oriented philosophy of the company.

Such close cooperation is something big customers are beginning to expect and demand. Thus, for instance, Ford plants in Germany expect the bearings factory in Airasca, Italy, to check the inventory control system at Ford twice a day to determine Ford’s needs for bearings, and make deliveries accordingly. Sales management at the factory assumes that most of their customers (mainly car manufacturers) will have similar requirements in the near future.

This trend towards more customer orientation is explicit in official SKF documents. Here is the 1996 annual report: "Sales people are organized into market segments around customers rather than products in order to offer best possible service, solving their problems rather than offering a product." One example of such customer orientation is the interaction between SKF and Volvo’s motor plant in Skövde, Sweden. In this factory, 1 500 motors are manufactured every day. The machines are using SKF bearings. SKF has developed a computerized monitoring system to enable Volvo to have exact on-line information about the status of the bearings. The machines are equipped with sensors communicating directly with the plant’s computer system and this system is connected to an SKF computer system, transferring the information to SKF. When a machine begins to malfunction, SKF can advise Volvo on what actions to take, providing a service rather than just selling a product.

This is only one example of a new and rapidly growing market for SKF in products and services for preventive maintenance. These services include everything from small portable equipment for collection of information to on-line systems continuously collecting information and automatically stopping processes when needed. In 1996, SKF installed one of the most advanced monitoring systems in the world at Star Petroleum Refinery in Thailand. The system monitors, and prevents breakdown of, all rotating equipment at the refinery like generators, turbines, gear cases and pumps. The market for such systems is rapidly growing in the paper industry. Many of the products for this market involves a large amount of services, setting up continuing cooperation and interaction between SKF personnel and the customer company.

Increasing customer orientation places new and different demands on SKF and on its personnel. From being experts on the production of bearings, SKF has to learn more and more about their customers and their business. At the European sales office in Turin, for example, more and more people are becoming engaged in direct customer interaction. A help desk is set up for customer advice on design, helping customers identify the bearings best fitting the requirements of the product under design and giving advice about how bearings requirements affect the design. And for the Italian customers, an organization is set up making sure that all customers are contacted by telephone or visit on a regular basis, and a detailed database about these customers is compiled.

Some of these changes may seem innocent and marginal, but together they add up to a radical change in the SKF company culture, changing its focus from production and distribution to customers and service. As of yet, this change is only hinted at, with occasional examples and experiments like the ones reported above, but already this new focus puts a strain on the traditional production orientation and the infrastructure that has served so well to support it.

Through its Common Systems strategy SKF has developed a corporate infrastructure including all major applications, not only basic communication services, but sales order systems, production planning systems, and the like. This strategy has been very successful. Over the years, the systems have been adapted to changing needs without dramatic crises. The information systems infrastructure has been aligned with business strategy and under management control. This approach has only been possible, however, because SKF has remained a traditional production company with a relatively narrow range of relatively simple products. While staying within the old production paradigm, SKF has avoided the requirements for a flexible organization and radical changes. SKF has changed its organization at slow speed, a speed the infrastructure could follow.

Now, in the late 1990s, SKF seems to be flirting with a more radical change, however, in the direction of a service and customer focused organization. Such a change is difficult to take slowly, involving as it does a more radical change of organizational paradigm, from a production organization to a more diversified and flexible structure. And the infrastructure has to follow. The current computing setup (mainframes, text based terminals) has to be replaced as well. The new infrastructure has to support the new organization. It will be more heterogeneous and it will have to be far more flexible. Like the organization itself, it has to be more customer oriented to support a closer customer interaction. The monitoring system installed at Volvo in Skövde and integrated with SKF’s own systems in Göteborg is one example. Such an infrastructure cannot be owned by SKF alone. It has to be based, as far as possible, on public infrastructures like the Internet. Such a heterogeneous, rapidly changing infrastructure based on publicly available services, can only to a limited degree be controlled by SKF.

SKF is also approaching the end of its computing paradigm because it is built on dying technology. Graphical interfaces, client server architectures and multimedia technology will be required. Such a change in technology will be a major challenge independently of new organizational requirements. But it will contribute to a speeding up of the changes already begun in the organizational culture.

The challenge is this: A production oriented company with industrial customers can operate on a global market with its own internal infrastructure, and it will do well to turn its company into one integrated production and distribution organization. The drawback of this approach is the normal dangers of centralization: creating a slow moving colossus with little local initiatives and little sensitivity to local variations. SKF seems to have handled these dangers well enough in the past, giving its factory engineers a substantial amount of autonomy. But that was when production was the focus.

A service oriented company, on the other hand, will have to be very different. Its operation on the global market is made possible by public infrastructures, and its focus is sales and services rather than production and distribution, which it presupposes are already well managed. Were SKF to move in this direction, it would have to create a networked sales organization, which rather than Common Systems use public systems. That sales organization could either be closely tied to production and distribution or it could operate at a distance from the factories, treating them as outsourced partners.

To create such a sales organization is no easy task. When companies try to get closer to their customers, providing services, anticipating needs, being more aggressive in their marketing and sales efforts, company boundaries will begin to dissolve. But those boundaries will dissolve for other reasons as well, when companies want to change by increasing personnel turnover, encouraging collaboration across company borders, relying more on consultants and independent operators, and so on. In all of this, there is a trend away from the self-sufficient, formally organized, closed factory, to a more nervous, fragmented company, opening up to the world surrounding it. It is very difficult to see how such a fragmented organization could be tolerated within the SKF family.

Engineer is still king at SKF, and it will not be easy to change that. But the profit margins for old industrial companies like SKF are slim. And more forceful automation of the production to cut costs, necessary to meet global competition, will not change this fact. In this companies like SKF resemble their poor cousins the textile factories. Those factories are now functioning as low profit outsourcing operations in the hands of high profit design, sales and marketing companies. Will the mechanical industry go the same way as the textile industry? If so, then SKF may very well be wise to begin setting up its own more customer oriented marketing, sales and service organization close to the market. The experiments with more active sales initiatives and customer services, seem extravagant, but maybe they are a first step in the right direction. Some of the ideas of the IT people at SKF (as summarized in "The Future IS study") seems much too trendy for a company like SKF—mobile computing, multimedia and virtual reality technology—but maybe they are right too? And maybe pressures from customers to cooperate more closely, if only by simple means like email and telephone, and by sharing information systems, eventually will bring forth a new, more service oriented SKF?

But then, on the other hand, there are of course good reasons for SKF to stick to its successful tradition and remain what it always has been: a powerful industrial company focusing on production and the quality of its products. As long as the major customers of SKF, such as the automotive industry, electrical companies, construction companies, paper mills, and so on, don’t change, it does not make sense for SKF to change.

Eventually it may very well be the case that the business focus in the world will shift from powerful industrial production companies like SKF to powerful sales, marketing, and financial operators. Certainly it is the case that within the automotive industry there is such a shift going on from the production end of the companies towards the design and sales end. But this does not mean that the production part of an automotive company will have to, or should, change. And SKF may very well remain a machine bureaucracy, producing high quality bearings. But in order to remain in power, it would have to make sure to control the sales of its bearings. And in order to do so it would have to turn its sales organization into a much more powerful and customer oriented operation and increase its power over production.

How easy would it be for SKF to change its sales and marketing organization into a network of networking organizations? Well, first it would have to accept the power shift from production to sales and services, without at the same time giving up its high standards of product quality. Perhaps, the only way it could do this would be by making the sales and services offices more independent, distinguishing them more clearly from the production divisions. Perhaps, they ought to be able to sell products from other companies? Perhaps, SKF should divide into two company structures, one for production and one for sales and services? If it is true that the IT revolution will make the old and powerful industrial production companies lose their power to networking sales and service companies, then SKF might do well to make sure to have a powerful company of the latter sort when the time comes for the power shift.

We are extremely grateful to Olof Berg of SKF for being such a great guide to his company and to the Swedish Transport and Communications Research Board for funding the Internet project (http://internet.informatik.gu.se) of which the research reported in this book is a part.