Mass customization allows companies to provide every customer a product made according to his or her preferred specifications, delivered for a mass of customers. Building on advanced information management technology and highly flexible computer-aided manufacturing (CAM) capacity, this approach enables a company to create a large variety (scope) of “ad-hoc” customized products. The interactive capabilities of the Internet, particularly Web 2.0, make configuring and ordering the self-designed product much more accessible to the public. Different methods for customization and (personalised) recommendation of products have been developed and implemented in recent years, but only the approach known as mass customization (MC)  actually allows a consumer to  order a self-designed product item. Yet, MC  has not been adopted by companies in many consumer markets so far and programmes initiated  often survive for just a few years. The main impediment has been in lowering the costs to levels compatible with mass production. It raises doubts that MC can become a viable business practice.

An online MC programme provides the consumers with an interactive Web-based configurator or MC toolkit application for choosing their preferred attribute specifications, guiding them through the self-design process step-by-step. Graphic-rich and user-friendly interfaces help to enhance the experience for consumers. The Internet offers two important capabilities that can smooth the whole MC process: (a) gathering the preferences data from customers in real-time, and (b) transferring the information to a company’s facility from anywhere a consumer operates the toolkit on a personal computer or a mobile device connected online.

The best early example of MC implementation is probably that of the Japanese National Bicycle Industrial Company (NBIC — owned by Panasonic) that allowed consumers to order ‘tailored’ bicycles. But that was already available before the age of Internet: measures to fit a pair of bicycle to a rider were taken on a specially built physical model. Among MC applications available to consumers through the Internet in the past and present we may mention for example:

• NikeID for designing sports footwear (running for over ten years),
• Levi’s Orignial Spin jeans for women (terminated),
• Chocri chocolate bars and pralines from Germany (a UK service is currently suspended),
• Reflect.com customized cosmetics (suspended),
• Blank Label self-designed and made-to-measure  dress shirts for men (based in Boston & Shanghai and operating for four years),
• Lego’s Create & Share programme incorporated an MC service called byMe (terminated in Jan. 2012) that allowed users to order a box with the parts-bricks for the model they personally designed with LEGO Digital Designer — the toolkit is still available,
• Dell’s customized personal computers (changed customization approach).

In order to derive practical utility from configuring a product consumers should arrive to the task with adequate knowledge in the product category, understanding the attributes and their consequences with regard to quality or performance, and knowing which ones are the more important. This is particularly relevant for attributes for which there is shared convention as to options or levels that predict higher quality as opposed to attributes of more aesthetic nature and preferences reliant on personal tastes. Consequently, consumers are expected to have well-defined preferences on those attributes. However, many and even most of the consumers have just low to moderate levels of knowledge in any product category (e.g., food, home appliances, technologically advanced digital products). Furthermore, it is recognised now that consumers often do not have clear and well-established preferences and they resort to constructing their preferences as they advance towards a purchase decision. That means, for instance, that low-knowledge consumers who use an MC toolkit but do not clearly know what they are looking for are more likely to be influenced by the content of attributes offered for customisation by the product configurator and its overall structure.

But there is additional complexity to consumer response in the context of customization because the condition stated above on preferences may not be sufficient. Itamar Simonson, professor of marketing at Stanford University, expands the discussion by proposing that in addition to (a) having stable and well-developed preferences, consumer response to customised offers also depends on (b) the level of ‘self-insight’ into their own preferences and own judgement of their clarity and stability. When using the aid of a recommendation agent, it suggests implications such as the ability of consumers to accurately and clearly articulate their preferences to others, correctly acknowledging the real drive to their choices (e.g., rational vs. aesthetic or affective), and properly identifying a product recommendation that fits well their preferences (1). Consumers whose state of preferences is low on both factors are especially likely to be swayed by the attributes a recommending agent chooses to emphasise. In the case of using a product designer toolkit in MC, the burden on the consumer seems even greater, more explicitly requiring him or her to accurately articulate his preferences and subsequently confirm that the outcome product one designed indeed matches what he or she wanted; a major cause for consumers to abandon before ordering is their evaluation that the outcome product’s utility is less than planned. Another important cause is frustration and ungratifying experiences while utilising a configurator to self-design the product.

Consumers differ in the type of attributes they would want to customize, the number of attributes desirable for customization and the number of options or levels to choose from — factors that influence the purchase likelihood of a customised product. Interestingly, more knowledgeable consumers have not been found to be more inclined to purchase a customized product. Some differences in preference for layout of configurtors have been found related to variation in knowledge. For example, the less knowledgeable consumers are those who actually desire a larger number of options to choose from on attributes of personal subjective taste, because they tend to learn their preference as they look through options; high-knowledge consumers need that less. But we also have to take into account what consumers believe they know, and consumers are often wrong in that assessment (‘knowledge miscalibration’). Thus, overconfident novices are those who particularly want the higher number of levels compared with experts not sure of themselves (2).

Companies that engaged mass customization have frequently chosen a rather simple solution to these concerns: the attributes they offer for customization are primarily aesthetic, related to visual appearance of the product and much less to its actual performance. There is an over-emphasis of personalised features (e.g., posting a label of the customer’s name or an image created by her or him). Companies also tend to constrain the set of customisable attributes and offer very few of them — this is done not just for avoiding too much complication for the users  but for themselves, to leave them with more control over technical aspects of product design and the cost of making the customized products. While this may serve well the less knowledgeable consumers, it gives the impression that this is not a serious enterprise, more like a game or a ‘marketing gimmick’, which seems to lead the more knowledgeable consumers to dismiss this option for purchasing products. Even less knowledgeable customers may be disenchanted by constraints imposed in the wrong places.  Configurators should combine different types of attributes for customization that allow customers influence both functional utility and hedonic benefits (pleasure) from their product.

Companies have turned to other techniques such as recommendation agents and search assistants that would help customers find the most appropriate product model for them. A recommendation online system first probes the consumer about her or his preferences through a series of questions and then offer a set of product recommendations rank-ordered according to their match with the consumer’s preferences. This method is distinguished from MC in that it selects product versions from the existing assortment of the company and does not create a product specifically for the customer. This kind of aid satisfies the preferred balance for some consumers between the levels of perceived control they get and perceived assortment available, but it also depends on their belief that the system is more capable than themselves to find a product that matches their preferences. This may further depend on the amount of information asked for and on the type of procedure used to collect preference information. A search assistant that is common in shopping websites helps to drill through the assortment of product versions in a category and narrow it down according to attribute criteria chosen by the shopper, thus screening a smaller set of plausible alternatives. However such an assistant, that does not make recommendations, cannot be truly said to offer customization if it does not make use of preference information  from the shopper to organise his or her resulting set in a more efficient way.

Obtaining a product personally designed by the consumer may endow him or her with special positive feelings, providing an important drive to participate in such an activity. The benefits from MC pertain to the experience of designing or configuring the ‘private’ product as well as the subsequent value of the outcome product to the owner. However, researchers Franke, Schreier and Kaiser identified an extra effect they called “I designed it myself” that describes the subjective value, and elevating feeling, that arises from the consumer’s notion that she or he took part in creating the product. They suggest that this effect signifies that consumers would be willing to pay a higher price for the self-designed product compared with a similar kind of product picked off-the-shelf. The effect is contingent on an underlying sense or feeling of accomplishment of the consumer in his or her contribution to the product (e.g., that the effort invested was worthwhile, proven competency, pride). The researchers corroborate this effect in a series of experiments in terms of increased willingness-to-pay for a self-designed product and further show that it depends on the sense of accomplishment but does not exclude the role that perceived value of the outcome product has when making the purchase decision (3).

Companies that develop and implement mass customization programmes should take special care of a number of aspects of the interface consumers have with the Web-based design toolkit to improve their experience and enhance their satisfaction through the process.

• First measure that may be taken is to create at least two versions of a configurator, one that would be more suitable for more proficient higher-knowledge customers and another for amateur lower-knowledge customers. More generally, it is advisable to give users a greater degree of flexibility in choosing the complexity of configuring the product that matches the level of difficulty they think they can handle. In other words, a firm may allow some control to users in choosing whether they wish to set only aesthetic properties (e.g., visual appearance) of the product or also selected functional attributes, how many attributes to configure, etc..  Additional measures can be to invite users to show their creativity in features of visual design (enhances the sense of contribution) and recommending options on functional features of the product.
• Second, a company may target customers who are already more inclined to participate in other types of collaborative activities of product design and development, seeking the feelings of accomplishment, challenge and also enjoyment from this type of engagement (e.g., tie them together as LEGO used to do in its Create & Share programme). These customers may be valuable advocates that bring more followers to MC.
• Third, a variety of aids should be applied to provide users with explanations, examples or illustrations of the options for configurations, warnings about attribute combinations that would not work well, and a graphic demonstration that helps the user to realise how the product builds up.

In spite of discouraging hurdles in the past decade, it would be wrong to conclude that mass customization could not grow and expand. Yet, some changes may have to occur in the future that make it more advantageous for both companies and consumers to exchange benefits of assortment with personal customization. It may also take more time to find out for which product types consumer preferences can be more usefully answered through MC. Nonetheless. 3D-printing and MC may complement and push forward the utilisation of each other, depending on the level of autonomy consumers wish to have in co-creating their products. Technology is most likely to keep advancing, making the self-design experience easier and more gratifying, but technology will not solve all issues at stake and it is vital to continue studying and experimenting to better understand the human-side of consumer expectations of, processing capacity, and response to MC programmes as well as the ensuing 3D-printing.

Ron Ventura, Ph.D. (Marketing)

References:

(1) Determinants of Customers’ Responses to Customized Offers: Conceptual Framework and Research Propositions, Itamar Simonson, 2005, Journal of Marketing, 65 (Jan.), 32-45.

(2) The Role of Idiosyncratic Attribute Evaluation in Mass Customization, Sanjay Puligadda, Rajdeep Grewal, Arvind Rangaswamy, and Frank R. Kardes, 2010, Journal of Consumer Psychology, 20 (3), 369-380

(3) The “I Designed It Myself” Effect in Mass Customization, Nikolaus Franke, Martin Schreier, and Ulrike Kaiser, 2010, Management Science, 56 (1), pp. 125-140.