The false promise of mass customization

Abstract

Build to order will be hugely expensive and challenging, and its benefits are uncertain. But there is a cheaper alternative. For more than two decades, mass customization has been the future of manufacturinguand for some manufacturers it probably always will be. On the face of it, mass customization is a remarkably attractive proposition for consumers and producers alike. Consumers get a reasonably priced, tailor-made product reflecting their personal selection of colors, features, functions, and styles. Producers, for their part, get to reduce their inventories and manufacturing-overhead costs, to eliminate waste in their supply chains, and to obtain more accurate information about demand. In short, a win-win proposition. Two relatively recent developments have given the prospects for mass customization a boost: first, the success enjoyed by Dell Computer and other high-tech companies that build products to order and, second, the emergence of the Internet. Dell's story is by now a familiar one. Over the phone or via the World Wide Web, customers select what they want from hundreds of different components to configure the computer of their choice, which Dell doesn't begin to build until it has the money for it in hand. The company has become the envy of manufacturers of all stripes--from auto-makers to toothpaste formulators--that produce to forecast and are thus burdened with sizable finished-goods and components inventories and lagging cash flow. The second development-the Internet-gives manufacturers a platform for taking orders from a mass audience for customized products, such as bicycles, clothes, cosmetics, shoes, and vitamins, at almost no cost. In the past, customization of this kind was handled by skilled but expensive salespeople closely interacting with customers. But mass customization isn't necessarily feasible for all goods. Assembling cars, for example, is more complex and difficult than building computers. Still, auto companies such as BMW, Ford Motor, and General Motors have high hopes for the build-to-order (BTO) approach, a variant of mass customization. [1] They are convinced that the costs they can wring from their business systems by switching to BTO are enormous, that the benefits to customers are numerous, and that the many difficult challenges they face in implementing BTO can be overcome. Ford and GM have already launched experiments with full-fledged BTO for customers in a couple of cities. Other auto companies, in North America and Europe, have said in their public pronouncements that they will move to broad-based BTO systems in two to three years; privately, they have been heard to say that the change may take five or six. Nonetheless, we remain doubtful about the benefits of BTO. The challenges of implementing it in the auto industry are daunting-particularly in North America-and it isn't clear that the economics will work. Moving from a mass-manufacturing (or push) system of production, which automakers have continually refined over the years, to a BTO (or pull) system would require numerous operational and organizational changes throughout the auto industry value chain (Exhibit 1). Yet the payoff from all this activity is unclear. Luxury-car buyers seem eager to specify their preferences in great detail. But it is still too early to tell whether mainstream customers want their vehicles built to order, let alone how much they would be willing to pay for this. Fortunately, automakers bent on a BTO future have a alternative that could expand the customer's choice without presenting the challenges posed by real BTO. Yet this alternative would still allow car companies to get much better information about demand than they have now. Today automakers have their hands full making products that customers want to buy. For mainstream volume builders, BTO may be a distraction that they can ill afford. Focusing time and assets on the lower-cost virtual alternative is a better idea, notwithstanding the enthusiasm of the business press for BTO. …

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 132 citations exist upstream.