Product data management (PDM) looks like the revolution that never was. Says Steve Massie of Desktop Engineering: ¡§Most people aren¡¦t using PDM¡¨. Which is either odd or, depending on your point of view, just as well. Odd because it¡¦s hard to see how most companies keep track of their drawing data ƒ{ which drawing goes with which product, which is the latest drawing version, and who is now using which drawing and for what ƒ{ without it. Just as well because PDM wasn¡¦t developed for the chain of separately-owned one-process operations which the Internet makes feasible and which are increasingly the pattern for today¡¦s manufacturing. PDM systems might be able to use the Internet, but they often assume the business models in use before the Internet arrived, with design, marketing and production departments within one organisation. Says Tom Leeson of SmarTeam, a US PDM supplier just setting up in the UK, ¡§When companies talk about 80 per cent of investment in a product being committed in the design phase, [what they mean is] ¡¥let¡¦s look at design, let¡¦s work on our core competancies and then outsource things¡¦.¡¨ Each time they do this, he says, they create one more separate link in the supply chain: ¡§And they¡¦ve discovered,¡¨ says Leeson, ¡§that, in terms of reviewing engineering changes, actually communicating with the supply chain through normal mail or electronic mail or low bandwidth modems means you then reach other challenges.¡¨ The other challenges are the need to react even more quickly along this extended supply chain than you did when all these functions were in house. Which brings us full circle. Because it also means you can¡¦t react as quickly as you need to without meticulous control of all the design data. The data may be needed in any of those far-flung supply links and you have to know who¡¦s doing what to any of it. How well you can do this often depends more on the organisation you work in than on the technology itself. How so? For one thing the scale of the job is far larger than implementing something like CAD or ERP on their own. Says Massie: ¡§You can speed up a task, which is what CAD does, but to speed up a process that information flows through [offers] equal benefits in terms of improved speed and accuracy.¡¨ Massie acknowledges that the Internet changes the way manufacturing companies operate, and that it can help them achieve unprecedented levels of speed, quality and cost: ¡§But to improve processes rather than tasks you have to understand your processes. You have to know what you want to achieve, and it¡¦s a lot more complicated to implement than just going out and buying a piece of task-based technology ƒ{ that just models a widget or whatever. [But] tackling processes rather than tasks is difficult for many people in manufacturing.¡¨ The difficulties are not just important for companies in, say, consumer products, who are dealing with a large continuous flow of new designs. Nokia, for example, produces a new phone a month. Many companies have a basic product which they need to modify, quickly, to keep pace with the market. Nor are the snags confined to the supply chain. In some organisations competitive pressures force collaboration among units with different histories which barely know each other. In these cases, the cultural and organisational issues dwarf any problems with the technology. Take MEI, the Winnersh, Berkshire, based MEI coin-handling subsidiary of Mars Electronics. MEI develops and makes coin mechanisms for everything from vending machines to municipal car parks: ¡§Anything that will take a coin,¡¨ says MEI CAD manager Steve Friend. Coca-Cola and Pepsi are big customers, as is the London Underground and many suppliers of pub gaming machines. These customers don¡¦t tell MEI to adopt PDM and integrate it with its CAD systems. But Friend does report a growing need to exchange CAD geometry between MEI and those making the vending machines, and the need for rapid design changes to accommodate new coin sets implies the control of design data that PDM can give you. Some MEI designs have been out in the field for 15 or 20 years, says Friend. In other cases operators want constant design improvements. Todays¡¦ machines give change, for example, and customers want the machines to keep giving change for as long as possible: ¡§If a coin-mech runs out of change, people will stop using it. They¡¦ll move to the one next door, or the one down the street,¡¨ says Friend. MEI operates on three sites. Its sister company in Westchester, Philadelphia, USA, is of a similar size and number of CAD seats; and another, smaller office in Geneva has seven seats. Their markets used to differ. US machines took banknotes, and everywhere else took coins. That division is breaking down. The Euro makes European collaboration more urgent than before, and all three markets will converge as smart cards take hold: ¡§There is a growing trend for us to collaborate on designs,¡¨ says Friend. Designers and engineers from all three sites are having far more collective influence on new projects than ever before, and a growing design team, now 24 seats at Winnersh, makes design tracking harder. So the three sites are looking at common CAD and PDM systems. Before Winnersh installed its Solidworks CAD system from Cadtek 18 months ago it used Hewlett-Packard¡¦s ME10 and ME30 systems. So did the US operation, but Geneva used AutoCAD. The PDM choice was HP offshoot CoCreate, which will now only be used for legacy data. New designs will use SmarTeam PDM, which Winnersh has been using for about six months, plus ConceptWorks, a teleconferencing system which allows MEI¡¦s designers and its outside contractors to share and modify design information over the Internet. MEI is using contractors to source up to 90 per cent of some new projects. Says Friend: ¡§In recent projects most tooling for the plastic parts is being handled in places like China, the States, and Mexico. So we see ConceptWorks being a great aid in communicating ideas on individual parts.¡¨ If MEI¡¦s toolmaker has a ¡§then you can both look at the same part on the screen and you can both know exactly what he¡¦s talking about. You can even go as far as making the change live and him agreeing that¡¦s how it should be.¡¨ That¡¦s another reason for choosing SolidWorks, Friend explains. Its ease of use means MEI can choose among contractors it knows and teach them SolidWorks, rather than restricting themselves to contractors who already know a harder-to-learn CAD system. The benefits of all this? On a new project, says Friend, the design cycle is about half as long as a similar one eight years ago. MEI is over the worst of the transition to a new system. For Desktop¡¦s Duncan Gilchrist the worst is yet to come. Gilchrist sees a similar view from a completely different industry. He manages IT systems for forklift truck maker Boss Trucks of Leighton Buzzard, Bedfordshire. Integrated CAD and PDM, he says: ¡§ensures that the information that the designers are using is the latest information. The only way that we can ensure that is by having a PDM system that gives you that functionality. You need to [ensure] that the data that you¡¦re then going to be checking, redesigning, improving the product is going to be there and you can rely on it... and everybody knows what the improvements will be.¡¨ An example is a current project to improve the design of control levers for the movement of the forklift mast and carriage. Boss has no current need to exchange data with its German parent, Jungheinrich. ¡§But that is due to change,¡¨ Gilchrist says, because there is an increasing amount of joint work between Jungheinrich in Hamburg and its operations in the UK. The group is now evaluating suppliers for group-wide CAD and PDM systems: ¡§There will have to be process changes,¡¨ says Gilchrist. ¡§It¡¦s not going to be easy.¡¨ At the moment Boss uses a combination of CADDS V CAD and Optegra PDM systems, both from Computervision, which Parametric Technology (PTC) swallowed up in 1997. Jungheinrich might choose PTC or it could move to Unigraphics or Catia. Again, the issue is as much cultural as technical. Hamburg too is an Optegra user ƒ{ but the two operations manage their data in completely different ways. Boss only adopted CAD in 1996 and was able to start from a clean sheet, which Gilchrist explains means ¡§using parametrics, and working in assemblies and producing product structures¡¨. Hamburg has been using CAD for 20 years, ¡§so a lot of their processes were already entrenched. Their users had a lot of experience in the old technology and it¡¦s far more difficult for them to take this new technology on.¡¨ If the group moved to a non-PTC system, Boss would have to ditch Optegra and, ¡§A lot of the work that we¡¦ve done on site would to a degree be lost.¡¨ Another issue is that Boss has integrated its PDM and modelling systems but customised the Optegra PDM to suit Boss¡¦s design and production processes. Boss now has to plan upgrades carefully: ¡§We can¡¦t just say we¡¦re going to upgrade from version 10 to 12, because the products are integrated. So whenever we need to do an upgrade it takes careful planning. And obviously,¡¨ Gilchrist adds, ¡§there is a cost involved.¡¨ But Gilchrist accepts that the pain is part of achieving the benefits these systems can bestow: ¡§It¡¦s easy to stick your head in the sand, and say ¡¥I¡¦m happy with what we¡¦re doing, it works¡¦. Yes, it does work. But if you want to become an international player, part of a big organisation, then you need to sacrifice some things in order to move forward.¡¨ Friend¡¦s conclusion is equally robust. For his company, the combination of CAD and PDM has worked ¡§very well, lately¡K I don¡¦t think we could live without the PDM. [But] nothing is ever problem free. On an installation of a similar size to ours, the biggest thing I would suggest is you would try and implement a PDM system at the same time as a CAD system. Rather than get so far down the line producing all of this CAD data and then finding that you need to backtrack and put it into a PDM system.¡¨