Anthony Youhas couldn't resist. He started answering questions on the Web -- how he felt about low-fat foods, marshmallows, and nuts. Next thing he knew, General Mills (GIS) whipped up a custom brand of cereal just for him. For about $20, the 24-year-old software engineer from Scotts Valley, Calif. received 18 single-serving bowls of breakfast grub with his own mix of oats and nuts. Unfortunately, the packages were cute but pricey. And worse, they tasted flatter than a bowl of soggy Cheerios. "On a pure flavor basis, it was just as easy to go to the store and get a box," Youhas says.
No matter how tempting customized products may sound, consumers don't always bite when it comes to buying them online. General Mills' mycereal.com, launched in 2000, lost its crunch in less than two years. Another high-profile experiment by Procter & Gamble's Millstone Coffee, called Personal Blends, dispatched its last bean in the summer of 2001. Neither generated enough revenue to keep the Web sites running.
The lessons? Although the Web can be an efficient way to attract consumers, it's costly to customize products. And where tastes are involved, cyberspace often strikes out.
P&G (PG) learned that the hard way. To make a go of it, Personal Blends would have had to generate around $100 million annually, according to P&G engineer Robert Piotrowski, who helped dream up the idea. Only at that level could the company justify investing more. Meeting those numbers meant charging $10 per 12 oz. bag -- about the same as a top Starbucks Coffee (SBUX) blend.
Even then, P&G would have had to grind out and sell 10 million bags of coffee a year over the Web. That's tough when you consider that the U.S. has only 1.3 million high-income households -- the likely buyers of costly online coffee. And less than 30% of those recently visited sites with such goods as cereal and coffee, estimates Forrester Research.
A MATTER OF TASTE. What went wrong? Producing customized coffee is expensive. Starbucks.com succeeds because it can use the same popular blends and premade packages it sells in stores. Millstone promised to blend java to buyer specifications and label each bag with the buyer's name. "It had whole new cost structure that couldn't be funded," says Ken Cassar, an analyst at Jupiter Research.
It's also hard to define tastes in words. Consider the difference between coffee and cosmetics. As its coffee test fizzled, P&G found that Reflect.com, a site selling beauty items such as personalized lipstick and facial wash, would jell. Questions about cosmetics -- do you have dry or oily skin, a pale or dark complexion? -- are more straightforward. Food and beverages are a different matter. Despite survey questions about liking dark or semi-sweet chocolate and mild or spicy salsas, consumers "can't define their taste so clearly that asking them to personalize it online makes any sense," Cassar says.
Without the consumer expertise to choose a perfectly blended coffee or bowl of cereal, Youhas says, such offerings are "nothing more than a novelty thing." They certainly don't generate enough snap, crackle, and pop to last online. By Roger O. Crockett in Chicago