Mrs. Fields' Secret Ingredient

Tweet

The first is that if a machine can do it, a machine should do it. "People," says Randy, "should do only that which people can do. It's demeaning for people to do what machines can do. . . . Can machines manage people? No. Machines have no feelie-touchies, none of that chemistry that flows between two people."

The other rule, the one that keeps the technological monster itself in check, is that the company will have but one data base. Everything -- cookie sales, payroll records, suppliers' invoices, inventory reports, utility charges -- goes into the same data base. And whatever anybody needs to know has to come out of it.

Don't enforce this rule, and, says Randy, "the next thing you know you have 48 different programs that can't talk to each other." Technology grown rampant.

Having a single data base means, first, that nobody has to waste time filing triplicate forms or answering the same questions twice. "We capture the data just once," says controller Quilter.

Second, it means that the system itself can do most of the rote work that people used to do. Take orders for chocolate, for instance. The computer gets the weekly inventory report. It already knows the sales projection. So let the computer order the chocolate chips. Give the store manager a copy of the order on his screen so he can correct any errors, but why take his time to generate the order when he's got better things to do -- like teaching someone to sell. Or, take it further. The machine generates the order. The supplier delivers the chips to the store and bills the corporate office. A clerk in the office now has to compare the order, the invoice, and what the store says it got. Do they all match? Yes. She tells the computer to write a check. The more stores you have, the more clerks it takes. Why not let the computer do the matching? In fact, if everything fits, why get people involved at all? Let people handle the exceptions. Now, the clerk, says MIS director Quinn, instead of a processor becomes a mini-controller, someone who uses his brain.

The ordering process doesn't happen that way yet at Mrs. Fields, although it probably will soon as Randy continues to press for more exception reporting. You can see where he's going with this concept.

"Eventually," he says, "even the anomolies become normal." The exceptions themselves, and a person's response to them, assume a pattern. Why not, says Randy, have the computer watch the person for a while? "Then the machine can say, 'I have found an anomoly. I've been watching you, and I think this is what you would do. Shall I do it for you, yes or no. If yes, I'll do it, follow up, and so on. If no, what do you want me to do?" It would work for the low-level function -- administering accounts payable, for instance. And it would work at higher levels as well. "If," Randy says, "I can ask the computer now where are we making the most money and where are we making the least and then make a decision about where not to build new stores, why shouldn't that sort of thing be on automatic pilot too? 'Based on performance,' it will say, 'we shouldn't be building any more stores in East Jibip. Want me to tell [real-estate manager] Mike [Murphy]?' We're six months away from being able to do that."

The ability to look at the company, which is what the data base really is, at a level of abstraction appropriate to the looker, is the third advantage of a single data base -- even if it never moves into artificial-intelligence functions. It means that Debbi Fields and Richard Lui are both looking at the same world, but in ways that are meaningful to each of them.

The hurdle to be overcome before you can use technology to its best advantage -- and that isn't equivalent to just hanging an outboard motor on a skyscraper, as Buckminster Fuller said -- isn't technical in the hardware sense. Randy buys only what he calls plain vanilla hardware. And it isn't financial. For all its relative sophistication in omputer systems, Mrs. Fields spends just 0.49% of sales on data processing, much of which is returned in higher productivity.

Much more important, Randy says, is having a consistent vision of what you want to accomplish with the technology. Which functions do you want to control? What do you want your organization chart to look like? In what ways do you want to leverage the CEO's vision? "Imagination. We imagine what it is we want," says Randy. "We aren't constrained by the limits of what technology can do. We just say, 'What does your day look like? What would you like it to look like?" He adds, "If you don't have your paradigm in mind, you have no way of knowing whether each little step is taking you closer to or further from your goal."

For instance, he inaugurated the daily store report with the opening of store number two in 1978. The important thing was the creation of the report -- which is the fundamental data-gathering activity in the company -- not its transmission mode. That can change, and has. First transmission was by Fax, then by telephone touch tone, and only recently by computer modem.

Having a consistent vision means, Randy says, that he could have described as far back as 1978, when he first began to create it, the system that exists today. But he doesn't mean the machines or how they're wired together. "MIS in this company," he says, "has always had to serve two masters. First, control. Rapid growth without control equals disaster. We needed to keep improving control over our stores. And second, information that leads to control also leads to better decision making. To the extent that the information is then provided to the store and field-management level, the decisions that are made there are better, and they are more easily made.

"That has been our consistent vision."

 PREV  1 | 2 | 3 | 4