The predictive power of Internet-based product concept testing using visual depiction and animation

One critical step in new product development is selecting from among multip le possible product concepts the one that the firm will carry forward into the marketplace. There is a need for low-cost, parallel testing of the appe al of new product concepts, the results of which closely mirror ultimate ma rket performance. In this article, the authors first describe an Internet-based product conce pt testing method they developed that incorporates virtual prototypes of ne w product concepts, substituting them for physical prototypes. The method c an be used with either static representations of the products or with dynam ic representations that demonstrate how the product works through a simulat ed video clip of its operation. The objective of this method is to allow de sign teams to select the best of several new concepts within a product cate gory with which to proceed without having to develop physical prototypes. The authors then provide a rigorous test of both virtual prototype methods against tests using both physical prototypes and attribute-only (i.e., no v isuals), full-profile conjoint analysis. Nine concepts compete against two actual products in rite rests. Market shares from the test using the physic al prototypes are defined as the "actual" market shares. Predicted market s hares for the attribute-only, full-profile conjoint analysis and each of th e two virtual prototype methods are compared to those obtained for the phys ical prototypes. Both static and animated virtual prototype rests produced market shares tha t closely mirrored those obtained with the physical products, outperforming the set of predictions across the full range of products produced in the a ttribute-only conjoint analysis, Interestingly, the attribute-only conjoint analysis identified the top three products, in correct order. It was unabl e to differentiate performance below these top three products. Furthermore, it predicted market shares for the top three products to be well below tho se achieved using physical prototypes. As virtual prototypes cost considerably less to build and test than their p hysical counterparts, design teams using Internet-based product concept res earch may be able to afford to explore a much larger number of concepts. Vi rtual prototypes and the testing methods associated with them may help redu ce the uncertainty and cost of new product introductions by allowing more i deas to be concept tested in parallel with target consumers. (C) 2000 Elsev ier Science Inc.