Act 2, Scene 2: Restructuring at the Factory

Act 2, Scene 2: Restructuring at the factory

The relationship between British and Japanese engineers and managers was rather different at the factory than in R&D. This can once again be placed in the context of the logic of the take-over for Mitsubishi and for Apricot. Mitsubishi was particularly dependent upon tapping the design and development expertise of the Apricot R&D team, and while the manufacture of PCs was also influenced by the distinctive character of the design and the marketing of computers Mitsubishi management saw it as closer to some of existing operations. Indeed Mitsubishi believed that not only their financing of investment but their production engineering skills should contribute to the refurbishment of the Scottish production facility, though they recognised that there was no simple analogue to be transferred from Japan. Thus the influence of Japanese managers and engineers was rather more evident and extensive at Glenrothes than in Birmingham, but it was still mediated through the distinctive features of Apricot’s PC operation.

The new SMT facility, which had been funded by the parent and involved processes which were common to many of their other factories, saw a heavy involvement of Japanese engineers in installing the production process, training the operators and later trouble-shooting various production problems, but even here local management emphasised that the continuing management of the work process was now in their hands. Meanwhile employees suggested that neither Japanese nor British managers had moved far to enhance worker flexibility and product quality in this area. Secondly, a major reorganisation of the main production area involved significant Japanese engineering input, but in a form that was also heavily influenced by sector recipes rather than mirroring established practice in Japan. This reorganisation involved a move to single-line cell-build (at the centre of which individual operators built-through whole units of a variety of models) from the older dedicated flow-lines, to provide the basis for a substantial growth of output which would also have to accommodate an extended model mix. The way in which the combined team of British and Japanese production engineers addressed this distinctive set of production requirements was strongly informed by a range of visits to other electronics factories within ‘Silicon Glen’.

Thirdly, Japanese concerns about product quality were influential in the way the production process was operated, not so much directly through the presence of Japanese staff at the factory but rather through the role of Mitsubishi Electric as a customer because one of it’s marketing businesses imposed stringent standards for the large volume of machines being produced for the Japanese market. This led to continuing negotiation over the balance of cost, durability and innovation appropriate for the PC market world-wide, and the resulting quality standards continued to met primarily by rigorous and repetitive testing procedures at the end of the line, and rectification, rather than a sophisticated enforcement of ‘building in quality’. Interestingly the senior Japanese engineering manager was rather more relaxed about this than the British quality manager, who bemoaned the inadequate commitment of production management to quality procedures (Quality Circles, for instance, fell into abeyance in the mid 1990s) in the face of the usual round of production pressures and crises.

Thus any drawing upon parent company production techniques was negotiated through a process of managerial micropolitics and alliance formation, in which arguments and understandings about the scope and limits of these techniques in particular social, organisational and technical settings were rehearsed, but this often involved a recognition that established approaches and sector recipes were of greatest relevance. While teams of Japanese engineers visited to advise on machine operation or layout or propose ways of coping with production or quality problems, this very rarely meant wholesale transposition of production arrangements.

A senior Japanese manager at Glenrothes, who had been key figure in engineering innovations at the factory, had worked on small business computers in Japan, when Mitsubishi Electric moved into that area from its mainframe business. On that basis he had been involved as an engineering specialist in the parent company’s assessment of Apricot before the take-over and after the purchase he was then despatched to become a senior Japanese manager there. He saw his specific remit as to draw upon the parent’ company’s manufacturing expertise:

Mainly my job was so we could get the infrastructure. It was difficult, [Apricot was] a very young company. From the manufacturing point of view not so strong, also comparing Mitsubishi Electric. At that time Mitsubishi Electric [had] already decided to import the Apricot product to the Japanese market and sell their product through Mitsubishi Electric channels. That means that the product should [must] be very reliable. That means Apricot needs to create good manufacturing process. [7]

This meant overcoming substantial differences between the approaches of the two companies. In particular, while Mitsubishi planned production in detail to guarantee quality, production planning at Apricot was less detailed and also more subject to revision. He felt that this difference made it difficult for Japanese and British managers to collaborate, though such co-operation was important. At the same time he acknowledged the significance of the specific engineering expertise which had been developed within Apricot’s PC business. However, as will be shown below, the implications of these features varied in different areas of the factory.

Direct transfer and its limits: the SMT story

So far as the SMT line was concerned, the initial investment decision was designed to address quality problems arising from distant production in Singapore, and was strongly led by Mitsubishi Electric

When you’ve got something like a week between when the board is manufactured and the fastest you can get it into the UK, your cost and flexibility start to suffer. So you end up with lots of work in progress ... lots of re-work ... lots of commuting backwards and forwards, all of which got very expensive and quite difficult to manage. ... The first thing they [Mitsubishi] said was ‘right we are a manufacturing company, we are going to have circuit board manufacturing next to where you assemble the computers. It’s all under control.’ They’d got a great deal of experience and expertise in this area and said that ‘OK we can support Apricot to reduce its costs and previous flexibility problems by making this investment’ so ... the planning started in [the year of the take-over] [4]

The introduction of an automated Surface Mount line at the factory, together with its associated manual insertion and testing operations, was intended to make the Scottish plant self-sufficient in PCB/motherboard manufacture. Since the parent company already had experience of operating such facilities in Japan this was seen as involving a more or less direct transfer of a full manufacturing process from Japan to the UK, with only minor modifications which related to such matters as shift working arrangements.

Senior British managers emphasised that, once the team of Japanese managers got the line up and running, it was then taken over by local management. Meanwhile, the experienced female workforce on that section saw the initial input of the Japanese engineers as important, for example as trainers in soldering techniques and in implementing new procedures for testing and rectification which had improved quality. However, they also emphasised that the older production pressures remained paramount. This underpinned scepticism about the way work was organised in SMT and elsewhere in the factory, because of the constant pressure and also sudden switches in production to meet priorities. One consequence of this was that moves between jobs were experienced as being shuffled around to meet production crises rather than as any form of rounded training:

That’s what’s wrong with this place I would say lack of training. .. I don’t know how many jobs I’ve done through there. I could be two or three weeks on one job and then I am pushed on to another job and I’m on that for a few weeks and then I’m off that. ‘Flexible [Name]’ that’s what it is. ... It’s like working on the line here. This company, quantity not quality and I always said that from the day I came in - its quantity. [8]

Hybrid innovation in production methods: the GEM line

While the SMT facility represented an addition to existing manufacturing capabilities, the other major area of manufacturing engineering innovation involved a radical reorganisation of the bulk of the existing operation. This sidelined the established range of separate and dedicated flow-lines, each involving a classic sequence of highly fragmented assembly tasks, and replaced them with one interlinked assembly process, organised in a sequence from kanban (kit-assembly), though ‘common build’, then ‘cell build’ and finally tests. A key Japanese engineer played the dominant role in the development of this new production line, together with some briefer visits by other Japanese engineers. Thus in explaining the distinctive production rationale of the new line a British production engineer emphasised that:

That idea came from the senior Japanese manager, for that method of manufacture ... The process that we do now I would say has been his baby ... We try to go for as much common elements as possible, like common elements within the system, and then what do we need to do to configure it for the customer - which is usually software, drive size and motherboard type, so that is how this thing has been set up. ... What we tended to be good at in the past [was] doing things in batches. Again, [the senior Japanese engineering manager] was very interested in being able to do it on a one-off basis. Because we do it as a one off thing, we do batch products as well ... We’ve also got material planning for a system which looks at a 10-day rolling forecast and looks at shortages of material. The idea is that this common assembly is good as a repeatable thing, the quality level should be repeatable [9]

At the same time managers at Apricot, including the lead Japanese engineer, also stressed that the GEM line had not involved a direct transfer of established manufacturing engineering solutions from Japan. Instead it was ‘our own concept’, original to the Apricot factory. It drew on ideas gleaned both from other Mitsubishi Electric factories and, more importantly, from visits to some of the computer companies which were in direct competition with Apricot in Europe. While access to the latter had not been easy, ‘in this country many engineers move to other companies so our people have some relationships’ and contacts had also been made through equipment vendors (compare Smith et al 1990).

Indeed, top management at Apricot emphasised the importance of these domestic and sectoral influences, rather than the specific expertise of Japanese advisers, despite the secrecy engendered by competitive rivalries:

We go off and visit ... where there is competitive plants ... Go and have a look at the American plant as to what is the most successful manufacturing recipe for PC industry and we’ve actually got those companies coming back to us now. We’ve had Compaq trying to get in to see our facility because we think we have now probably got the most flexible semi-automated facility going today. ... Openness is what the PC industry is about. What does that mean? It means relationships with as many people as is required to get the job done. Don’t believe you know anything any better than anybody else. Copy wherever you can. Utilise and plagiarise is really the only thing we should be about in the PC Industry. [10]

Given the differences between the products and markets of most of the Japanese parent factories and those of Apricot, it was perhaps unsurprising that the senior Japanese engineer had spent over two years investigating the requirements of PC production, trying out ideas on a small product line and planning and developing the GEM line itself. In the end this resulted in a hybrid production line which sought to embody three key features:

(i) contrasting labour requirements and work tasks along the single line, with short-cycle routinised tasks on the early ‘common build’ part of the line, and longer cycle (initially 12 minutes) more varied partial build-through and testing routines in the later ‘cell build’ area, where individual operators had the responsibility to complete the assembly by adding the customised elements onto each machine to order;

(ii) this avoided the persistent line-balancing problems of building small batches on conventional assembly lines by taking the more complex build throughs temporarily off-line, but the maximisation of through put with such varied production relied heavily on tight work scheduling and effective quality control on the single line;

(iii) implementation of a version of a ‘kanban’ supply system, based on holding a day’s stock of most components, though the variety of models and the power of dominant suppliers to ration vital high value parts made a classical internal kanban system designed for mass production inappropriate.

This innovation had major consequences both for the relationship with R&D and for the character of employment relations in the factory. Firstly the universal line placed a premium on ‘design for manufacturing’ so as to maximise the scope for ‘common build’, and this increased the leverage of the manufacturing engineers in pushing DFM up the design process, with the backing of Apricot’s top management (‘The factory has to be more and more powerful, say I am sorry, we are not going to build that product if you design it like that’). Secondly, factory managers interpreted the contrasting labour requirements along the new line as an opportunity to differentiate the workforce more systematically into permanent and temporary workers: