(FORTUNE Magazine) – Despite his title of chief administrative officer, Anthony Rucci functions as one of a handful of true chief learning officers in the world, charged with revitalizing Sears Roebuck's dinosaur culture. His peculiar job description includes responsibility for HR and legal, as well as an initiative to cut costs on $15 billion of goods and services that the retailer buys annually--and for raising the consciousness of the work force by sharing valuable information. To do that he relies on customer feedback to change employee behavior and improve performance. His assignment has been to help a work force of about 300,000, mostly sales clerks and blue-collar workers, learn to focus on customers. So far, so good: Total return to shareholders has averaged over 30%, compounded, for five years. In a conversation with FORTUNE's Stratford Sherman, Rucci argues that the fundamental challenge for Sears Roebuck, and for any company in any industry, is to help its employees understand the business and its customers.

What does a chief learning officer do?

I don't like that title because it suggests the opposite of what we want, that learning will emanate out of one executive's office. The learning function is about giving people the information they need to have an enlightened opinion about how to do their job better. During my first 90 days here, I visited 40 or 50 of our stores, meeting with small groups of hourly associates and managers. I would always ask, "How much profit do you suppose Sears keeps on every dollar of revenue?" The median response was 45 cents after tax. At the time we were actually earning 1.7 cents on the dollar. So you'd hear that and go holy cow, we have an economic-literacy problem here. That's management's fault. The second thing I'd ask is "What do you think is the primary thing you get paid to do here?" Over half the people answered, "To protect the assets of the company." In the early 1990s, Sears Roebuck got some of the lowest customer-satisfaction scores in the retail industry--surprise, surprise.

So where did you start?

First we had to return to profitability in retailing. Then, in 1994, we grouped the top 120 executives into task forces on themes, such as customers, employees, and financial performance. Their task was to define world class in each area and give us measures to track progress. Those task forces had an amazing impact. They helped us arrive at our vision for the business, which is to become "a compelling place to shop, work, and invest." We piloted 14 of their proposed measurements and eventually selected six to build our performance index: one measure for compelling place to work; two for compelling place to shop; three for compelling place to invest. The resulting data led us to formulate a theory that said financial performance is a lagging indicator--by the time your company publishes its annual report, it's too late to do anything about it. But there are some leading indicators that predict what financial performance will be. In our case they turn out to be things like employee attitudes and whether customers see our stores as fun places to shop.

Isn't that obvious?

It's a blinding flash of the obvious. But it's amazing how many major U.S. corporations, including Sears, lost sight of the importance of their customers and their employees. What gets measured gets done. We knew that unless we produced credible, auditable measurements in all three areas--shop, work, and invest--all the attention would gravitate to financials, and we wouldn't get the traction we need on shop and work.

This is pretty soft stuff you're talking about, quality of work life and customer satisfaction. How do you make the measurements rigorous enough to be credible?

We went to Claes Fornell International Group, an organization made up of econometric statisticians at the University of Michigan, and said, "Look, we have this theory about what drives our business. Can you test it?" We provided 13 financial measures for all 820 of our full-line department stores, hundreds of thousands of employee-satisfaction data points, and millions of data points on customer satisfaction. On compelling place to work, for example, we had collected years and years of data from a 68-question survey that went out to all employees every two years. CFI took all our data and put it through their causal modeling technique, which--this is for statistical jocks--goes beyond simple correlations and allows you to make inferences of cause and effect with some degree of predictability.

CFI told us that employees' attitudes about the job and about the company are the two factors that predict their behavior in front of the customer, which in turn predicts the likelihood of customer retention and customers' recommending us to others, the two factors that, in turn, predict financial performance. The empirical data ended up exactly replicating what our task forces had told us two years earlier. Now we know that if a store increases its employee-satisfaction score by five measuring units this quarter, the following quarter its customer-satisfaction scores will go up by two units. And if a store increases its customer satisfaction by two units this quarter, its revenue growth the following quarter will beat our stores' national average by 0.5%. It's not guesswork or theory anymore. We have built an empirical model that says unless you have a trained, literate, motivated, competent work force, and give them decision-making authority, you don't get satisfied customers no matter how good the merchandise is. The right work force creates customer satisfaction, and that produces superior financial performance. We have been tracking this for two years, and I can tell you this prediction is absolutely holding up.

How does this understanding affect people's compensation?

We've only been at it a year and a half; we're still in a learning mode. Our top 200 executives, 120 of whom were involved in building the six-measurement model we now use, have 30% to 70% of their incentive compensation driven off of it. Eventually, employees at every level will have some of their pay linked to nonfinancial performance measures. We have already launched a much more comprehensive customer-survey program that will allow us to rate individual associates on the customer satisfaction they produce. In March, cash registers in nearly all our retail stores began randomly printing out a special coupon along with customer receipts in a small percentage of transactions. With 468 million transactions every year at Sears' registers, it's not hard to generate a statistically reliable sample, even for an individual sales associate. The coupon says that if you agree to call our 800 number in the next week and answer prerecorded questions by pressing buttons on your touch-tone phone, we'll give you $5 off on your next purchase at Sears. Because each call is linked to a particular transaction, we're collecting meaningful data on individual sales associates for the first time. Ultimately, we're going to be able to go to Mary Smith and say, "Mary, you waited on 5,000 customers in the last three months; 250 of them were asked to provide data and 70 of them did. And here is how they rated you personally, Mary, on service, friendliness, product knowledge, and so on." But before you get to a new compensation formula for those sales associates, you have to raise economic literacy by giving them business information. Then people in the stores have to participate in whatever goal-setting process you've set up. That's how you promote ownership.

Let's talk about economic literacy.

That's a theme that weaves through everything we have done. Economic literacy here didn't mean putting everybody through training in finance. We wanted it at an operational level. The main communication device we have used is town meetings built around a series of four learning maps we produced with Root Learning of Perrysburg, Ohio. They boil down a huge amount of information into a visual format that makes it very easy to absorb. These maps are six feet by four feet, so there is room to pack in a lot of data in the form of charts and graphs, which are integrated into scenes with familiar images and a very basic story line, such as a car driving past some stores. The first one, called "A New Day on Retail Street," shows how the competitive environment has changed since the 1950s--for instance, that the frequency of consumer trips to shopping malls declined by two-thirds between 1985 and 1995. We had a forklift operator in Detroit who looks at this and says, "Wait a minute--if people are going to the mall one-third as often, and all our stores are in malls, why are we spending so much money remodeling those stores?" I'm sitting in the back of the room going hallelujah! You want people to know enough about the business to ask those kinds of questions.

The second map in the series was "The Voice of Our Customer." It showed how we compared on overall customer satisfaction with our major competitors--and the news was very bad. Many of our associates were shocked. And they realized something had to change. Access to information is what motivates change and improvement. It was just as essential that everyone understand that Sears isn't making 45 cents on the dollar after tax--now it's more like 3.3 cents at our recent peak.

Sears has made great progress by a number of your key measures, particularly customer satisfaction. But how are your ratings from employees?

Our scores have improved, but I would say that our employee-satisfaction scores, measured against a world-class standard, are just okay. During the last 18 months the score rose two points, to 68 on a scale of 100. Could we drive this score to 90? None of us know. No benchmarks exist.

But the ultimate test is how we perform in front of the customer, and according to FORTUNE's latest survey of customer satisfaction [Feb. 3], Sears Roebuck produced the fifth-highest improvement among 206 companies. If you believe our econometric model, it's our employees at the point of sale who drove that.

In the U.S., retail stores typically have employee turnover of nearly 75% a year. If your turnover is anywhere near that level, how can Sears' learning efforts ever make a lasting impact?

It's a chicken-and-egg issue. Our econometric model tells us that if we continue to improve our scores on employee satisfaction, turnover will decline. Improving those scores is job No. 1 for the next year. We have to keep repeating the basics like learning-map sessions while getting better at empowering our people. A lot of that is just getting us managers out of their way. When people get a chance to accomplish something themselves, they build self-esteem. They just come to life.

Like watering seeds?

Exactly.