The dreaded time has come to buy that new technology gadget you've heard so much about. You know in general what you want, but with so many brands offering such a confusing range of features, how is even a savvy consumer to decide? And how can the manufacturer best appeal to such puzzled consumers?
A team of Sloan School marketing professors and students has come up with a promising, web-based answer for company and consumer alike.
The Virtual Customer Initiative (VCI), which combines complex computer calculations, multimedia tools and web interactivity to quickly and accurately obtain customer feedback, is already available online for free. And a spinoff from the VCI effort, the Intelligent Advice Module, holds real promise for a PalmPilot-type wireless device into which users can program personal shopping preferences to help them make easier and better-informed purchasing decisions.
"When we started VCI, our goal was to help firms determine what consumers want in a way that's faster and more accurate than traditional market research tools, while giving people an incentive to provide accurate responses," said Sloan Professor Ely Dahan.
A typical market research project today can take up to six weeks and cost up to $200,000. VCI cuts that time substantially and yields results that are as accurate as those obtained through traditional means.
VCI utilizes six web-based methods (the number could grow to 20 or more) to pinpoint customer preferences. The first set of tools, based on computer-driven mathematical formulas developed by the Sloan team, determines customer attributes such as desired features and price range for a given product. The other tools link those attributes to product designs that are then shown to the consumer. Through this web-based interaction, product winners emerge.
"Research to date indicates that these web-based surveys are proving accurate," said Dahan. "Our results closely predict consumer choice."
As both a test and an example of VCI potential, the team surveyed six different groups of people totaling a few hundred about their preferences among eight new vehicles, all of which were about to come online in a particular market niche.
"We were able to predict very accurately which ones people would like and hate," Dahan said. "The vehicles that our research indicated would be a disaster have in fact been disasters. Combine the computer and the web and you have a powerful tool for market research. We can use six or seven questions to achieve results that might once have required 30 questions."
Another example of VCI in real-world practice involves Polaroid's I-Zone instant camera, a highly successful product targeted at a younger audience. Web-based surveys of camera-buyer preferences revealed huge interest in a product feature--changeable covers--that Polaroid's development team hadn't considered so important. Changeable covers were added, and have helped fuel I-Zone's extraordinary sales success, especially among young buyers.
Why, then, is Sloan offering such a powerful marketing tool for free to anyone who wants to download its source codes?
"We're giving it away because we believe that that if companies get very efficient at using VCI, the number of product concepts that get evaluated goes up, along with the quality of the winners of that process," Dahan said. "You end up with better products that fit people's needs better because more products were tested at a lower cost, and much more quickly."
So far, the VCI web site has received more than 3,500 hits since going online in January, a significant number given the lack of publicity for VCI. The web site serves as a template, much like Excel or PowerPoint, said Dahan, who added that feedback from companies has been very positive.
Excited as the Sloan team is about VCI's potential to help firms, members are just as convinced about the potential of VCI methods to help individual consumers make informed, on-the-spot product decisions using wireless communications Intelligent Advice Module (IAM) devices.
Just as companies use VCI's computational formulas and web-based interaction to determine consumer preferences, the portable IAM runs through stored formulas that combine an individual's prior product choices with current preferences and other data to produce a decision that matches the owner's priorities. Similar research efforts aimed at developing smart shopping agents have been underway at MIT's Media Lab.
Dahan used a personal example to demonstrate IAM's potential. He was hired by MIT in 1998 but had a hard time finding suitable housing in a tight real estate market. "I kept striking out," he said. "So I ended up doing a VCI-style interview with myself. I took all the attributes I wanted in a house--good schools, proximity to MIT and so forth--and used conjoint analysis to prioritize them. It turned out that the amount of land with the house was a low priority, so I ended up buying a house with no land."
"As a consumer, what excites me is that just by storing information, this technology could make me a lot more efficient in my search for things to buy and do."
One cloud hangs over both VCI and IAM: how to record personal preferences and financial information in a way that makes such data useable, but which at the same time protects personal privacy.
"For VCI and IAM to be effective, people should have the incentive to be accurate and truthful when revealing their personal preferences," Dahan said. "The challenge is to develop encryption, biometric safeguards and other privacy protections so that people can securely own their personal preference information, have it work on their behalf and still protect their privacy."
The Sloan VCI research team is led by Kirin Professor of Marketing John Hauser. In addition to Dahan, the team includes marketing professors Duncan Simester and Drazen Prelec. Professors Andrew Lo and Robert Freund of Sloan and Tomaso Poggio of brain and cognitive sciences also contributed to the VCI methods. The initiative is funded by the Center for Innovation in Product Development and the Center for eBusiness@MIT at the Sloan School.
A version of this article appeared in MIT Tech Talk on October 24, 2001.