The future of…

Click here for contact details, other articles and personal details

The future of manufacturing

Ian Pearson, BTexact Technologies

March 2003

It may look at first sight that manufacturing would be relatively immune from internet impacts, dealing as it does with physical matter and processes rather primarily information. It is certainly true that a car can't be assembled on a web site and downloaded to a customer via the net. However, products will increasingly include information technology, and while manufacturing equipment might be fixed at a specific geographic location, even that may be accessed and controlled from anywhere via the net. Of course, the non-physical departments of a manufacturing company are as open to change as any industry sector. Let's look at the whole product life cycle to see what changes are likely.

Starting with the invention process, computers will become increasingly adept at analysing customer needs, based on the enormous volume of data collected at the point of sale. Trends can be spotted and linked to lifestyles, and gaps in the market identified. Computers are also already inventing useful objects, by taking ideas that have been developed in one field and applying them in another. This cross fertilisation is often used by humans when brainstorming but works just as well for machines. Humans and machines will mostly work together though, making the most of their relative strengths. Customers may be involved frequently during the later invention stages and throughout the design process. Even when a concept is just a few bits in a computer file, visualisation technology can help customers to appreciate the possibilities and feed their comments into the design process, often inventing features that perhaps an engineer might have overlooked and omitted. Virtual environment technology is able to use avatars (3d computer likenesses of people) in realistic synthetic environments to show users how a product might fit into their lives, thus taking a much more active part in the design and customisation of products. A woman choosing a design for a new wedding outfit might be able to try it out at a virtual reception with her friends, who are simultaneously co-ordinating their outfits for the same wedding.

People often resist the notion that computers can be creative, but they can already write music and novels that are almost indistinguishable from human produce. Design lends itself to similar automation. Computers can sample various design styles and produce endless variations and combinations. Evolution techniques can interact with the user to steer designs in directions that the user wants, thus increasing the degree of fit and improving personalisation and individuality.

Manufacturing itself relies increasingly on computer-controlled tools, and as they become more sophisticated, so we will see more personalisation and greater customer involvement. Since the net allows this interaction from anywhere, customer location is not an issue. Designers may be in a very different location from the fabrication unit that builds their prototypes.

There will certainly be a great deal more inclusion of information technology in products. Pervasive computing will include myriads of tags and activators that are built into products of all kinds. Radio frequency identity tags are the first generation of these, and are expected to replace barcodes in many products over the net few years. However, ultra-simple computing, data storage and sensor devices will also be developed. Products can be located, identified, and communicated with, even involved in processing. Chips will be absolutely everywhere. We are already considering many applications for chips inside our skin, and jelly babies that scream when you bite off their heads! However, adding chips doesn't imply adding huge expense. The same chip that powered the Apple 2 computer in the early 1980s would be 0.15mm across today, cost a penny and still run 2000 times faster! Identity chips are also expected to enter the market at around the penny mark. They can be interrogated and programmed from up to a metre away. Because these activators will generally be self-organised and use short range communications, they can be easily located in any product without the need for accurate positioning. We would expect them to be mixed in with the plastic in mouldings for example.

These chips will help a great deal with the trend towards treating some physical products as part of a service rather than being sold outright. The customer might buy a television 'service', and the actual TV set is remotely maintained by the leasing company. This allows service bundles to be tailored for long after the TV delivery. Adding activators gives the capability to include low cost processing, remote sensing, location and identification facilities. The service provider could ensure that the product only works at the agreed location, remotely diagnose faults, upgrade software and adjust the service levels. This kind of package might be suitable for many of the products we use in the future, and make economically feasible many services that rely on kit that becomes rapidly obsolescent. Many other products will still be bought outright, but in many cases will still contain such chippery because it allows the buyer to maintain them. Kids could police their belongings to prevent their 'borrowing', use or abuse by siblings.

Packaging will be affected by the changes in the way people shop. If they are dealing direct with the manufacturer, there is much less need for any packaging that is designed to catch the eye on a store shelf. It can be much more functional. This would reduce costs for at least part of the output. But it also will change the brand perception, which will become less visually anchored and more identified by advertising that links products to lifestyle aspirations. We will see a migration of brand influence also to third party quality assessors, who will help customer decide which product they want from the hundreds available worldwide from web based ordering. These assessors will tell you which one you need for your personal lifestyle and your budget.

The distribution industry takes the product from the factory gate to the customer, and this too will change. When dealing via the phone or net, we rely on postal services or couriers to bring the product to us, and often we are away from home when it arrives. This is obviously unsatisfactory and won't be accepted in a few years. Instead, we will have two tier distribution that takes the product from the manufacturer to a local warehouse, specified by the customer, and they act as our local receiver. They hold onto the goods until it is a convenient time for us. Over time, this local distributor may become much more powerful. Often we want goods of a particular specification, but don't care much who makes it, since quality is ubiquitously acceptable. We may be happy to ask our friendly distributor to use one of their contacts to get us what we need at a good price at the right time. They will be in a good position to develop and exploit this kind of loyalty and trust.

After-sales activity can be facilitated b greatly by means of chips in the product. They may be able to put an electronic manual on our TV screen via the web, or assist in fault diagnosis. Customers will be able to add new facilities or simplify interfaces by removing those they don't want. Products will slip easily into electronically managed homes and offices, tough we shouldn't expect these to be quite so widespread as some hype would suggest. In particular, we must guard against adding any functionality that isn't fail-safe or doesn't have an easy physical operation mechanism when the software inevitably fails. An automated kitchen sounds nice, but if it doesn't work when the kids need breakfast before going to school inn ten minutes, we will see a new social phenomenon of kitchen rage.

Finally, all products eventually die or need replaced, and green culture dictates that they should be recyclable. Clever use of materials has already given us phones that dismantle themselves when they are dropped into hot water. Embedded identification chips will direct materials for appropriate recycling too.

Apart from all these changes in the manufacturing cycle, we will also see the ubiquitous changes in corporate structures. Personnel and finance departments are routinely being outsourced already, and often automated at the same time, so that they need far fewer staff even in the outsourcing company. Other departments will see similar downsizing until we get to very lean companies with a few elite people backed up by an army of smart machines and software. The rest of the people will simply have to retrain, and because new jobs eventually get automated too, this will be a regular cycle for many workers. It is inevitable in the endless pursuit of better quality and functionality at ever-lower cost.

But we shouldn't be too concerned by this. If we could suddenly automate every job in the country, GDP would rise dramatically as we have all the machines doing all the old work, and still all the human labour available to do even more. The country will therefore be much richer in the future, and provided we can develop a socially acceptable wealth distribution, the future looks much better than today. Squabbling amongst ourselves about who gets the biggest slice of the cake will just slow down the benefits for everybody.

However, if people really don't like the way that industry is going, there is a final warning. The net gives future pressure groups the means to co-ordinate the actions of very large groups of people instantly, in ways that just aren't possible today. Consumer power will be very much greater, and any company that acts in a way that creates public anger will find the consequences much greater and much more immediate. Manufacturing industry has a great future, but it is one that is still intricately woven into the wants, needs and expectations of society. It will much less be able to ignore public opinion.