Directed decentralization: the Frito Lay story.

Back in 1930, our founder, Herman Lay, did everything. He bought and cooked the potatoes, packaged the chips and put them in his truck, brought them to the stores, and sold them. He did his own quality control: if people yelled at him when he went back the next day, he knew his chips didn't taste good. He had his accounts payable in one pocket and his accounts receivable in the other, and he could tell just by patting his pockets if he could afford to buy a few more potatoes or maybe a newer, bigger truck.

Herman was in constant touch with the marketplace. And so the company grew, took in partners, and built a group of strong regional companies, each of which put out products different from all of the others.

And then, in the 1960s, things began to change. Frito Lay merged with PepsiCo and, like other large companies at the time, emphasized its size and scale. it entered the era of national products, TV commercials, and professional management.

Built on functional excellence and high productivity, supported by national consumer polls, the company prospered. But, in the early 1980s, the competitive environment changed again. Small, regional companies geared to local tastes began to emerge, and large companies, especially those that could not move quickly, began to have problems. Because their frame of reference was national, the large companies were not focused on regional marketplaces and on the delivery outlets, or channels, for their products.

But advances in technology now make it possible for large companies like Frito Lay to return to the regional approach, to be in close touch with their customers, and to let people at every level of their organizations understand the dynamics of what they do and how what they do affects their companies' balance sheets. And they are able to do this while maintaining consistency in the quality of their products and the marketing leverage of well-known brand names.

Here's how we're doing it at Frito Lay.

Business volume and velocity One of the two most critical aspects of Frito Lay's business is volume. We sell four to five billion bags of our various products each year and move them from 40 plants through 10,000 salespeople into 400,000 stores. We had to master volume early on.

The second element critical to our business is velocity. Since we sell impulse items, we need very high service levels and penetration. If you walk by a shelf and our product is there, you'll buy it. But if it's not there, you won't go elsewhere and look for it. Furthermore, our product has a shelf life of 35 days, so we build very small warehouses. We invented "just-in-time" manufacturing because, if nothing left the warehouses for three days, our storage facilities would be full. At the other extreme, if production were to stop, the entire system would be empty in 10 days.

Complexity

Up to the early 1980s, our business wasn't very complex. We had national patterns. We were predictable. A bag of Doritos bought in New York tasted exactly the same as a bag of Doritos in California. We had an orderly process of building and implementing our annual plan, measuring its results, analyzing what worked and what didn't, and making the requisite changes. We had 10,000 employees in 40 plants who did their jobs very well.

Then the world changed, and, as our regional competitors got stronger, we kept introducing new marketing programs in the belief that we had to become less predictable if we were to be competitive. The problem was, as we became less predictable to our competitors, we also became less predictable to our own plants and sales force. So we had lots of short runs, shipped product by air quite often, and ran promotions on Thursday for a product that didn't arrive until the following Monday. As we shrank our cycle time, our business system became increasingly dysfunctional.

Systematic changes

To become flexible and sensitive to the local marketplaces while continuing to maintain national standards of quality and service meant we had to make changes to our business culture, organization, and infrastructure. We had to get off all paper systems, off the functional orientation, off the annual planning cycle. We needed to be a "shortcycle" company, and so we embarked on a seven-year program to completely rewire our information system. And, because our budget was essentially flat, we had to fund this project largely through the greater efficiency it produced.

we decided to build our new information system around three major frameworks: a planning and analysis system, an operational transactions system, and an executive decision support system.

The planning system is really a system for measurement. These measurements give us the chance to think about our company as sales, the same way small regional companies think about their businesses.

The planning system contains detailed data from each locality: its P&L, products, and promotion and pricing proposals. We also buy sales records from supermarket scanners, which contain information about our own products as well as those of our competitors. With this information, we can spot trends; we can begin to plan.

These data can be recalled by locality, by brand, or by distribution channel. All our field, brand, and channel managers have instant access to these data; if the competition appears to be gaining ground, our managers can immediately plan the appropriate promotion or pricing strategies. And, because the system tells everyone what is going on, our people in purchasing, manufacturing, and logistics can speedily implement any such plan.

In order to facilitate all this interaction, we built a lot of different technologies into the system so that users don't have to hunt for product codes or location numbers; nor do they need more than basic computer skills.

The transactional system

With our regional products, specialty brands, and the promotions we run at different times in different cities, our business has become very complex. If sales and operations are not in sync with each other, the plants will overship or undership products because they have to guess about demand, and then they end up airshipping or sending trucks out half empty. But because every one of our 10,000 salespeople has a hand-held computer through which he or she processes orders, we are able to coordinate sales and production effectively.

The goal of the transactional system is a constantly updated database that covers every inventory point. It will show what shipments left the plants yesterday and that today's shipments will be delivered tomorrow. Everyone who needs to, knows exactly where every bag of Doritos is. In due course, we will try to tie our suppliers into this network also.

Despite our short cycle time, this network of databases will allow us to anticipate sales and to meet our sales orders 99.5 percent of the time without having to do short runs at the plants. An additional benefit of the system is that all the actuals get posted into each region, channel, and brand where they belong-every bag, every check request, every gallon of gas that comes through the operational systems. Having these data in one place, rather than scattered throughout the company, gives us an immediate readout, allowing us to build forecasts.

The executive decision support system It's not unusual these days for a large company to install an executive support system. Usually, such a system is available only to the 20 or so top executives. What is unusual about our system is its pervasiveness throughout the company.

In this, the fourth year of our systems project, 120 people are linked in through PCs, but we expect to reach our intended total of 600 in a year or so-our senior staff and field managers at all levels nationwide. We call it directed decentralization, where decisions at every level can be made quickly and from a well-informed perspective. Everybody sees immediately where he or she stands.

Much of the project so far has been involved in getting our measurement system in place, which we have done. With our old system, we used to spend our time measuring things, but we didn't have time to analyze our findings. Now our measurement is done in a day, and we have the time to find out what all this information means. Why are we losing market share? Where did our plans go off, by how much, and who.

The system has also shown us how to eliminate unproductive effort. We haven't been forced to lay off people; rather, we have converted unproductivity into value-added activity.

We are also building analytical capability into the system, such as the set of rules that watches the competitive data. Our system software can, for example, examine the data loaded into our system from the supermarket scanners and send off an alarm when it picks up a new product. if the system picks up a new product in, for instance, St. Louis, it alerts our division manager in St. Louis, the brand manager of our competing product, and our senior staff. We track the success of the new product, and we can begin to develop our own new product if we feel it's necessary.

And so we're back to Herman Lay again. He knew about a change in his market as soon as it happened. With our new system, so do we. Information quickly moves up in the organization, and just as quickly down. Clear and immediate information is available at every level, and communication is constant and continuous. The world will never be simple again. Nor will it be predictable.