(FORTUNE Magazine) – THE UNION PACIFIC Railroad, the granddaddy of the American industry, the co- driver of the Golden Spike at Promontory, Utah, was run for a century with all the rigor of a model-train set. Blessed with wonderful routes that stretched from the factories of the Mississippi through the grainfields of the Midwest to the mother lodes of mountain ore and the groves and fertile valleys of the Pacific Coast, the railroad minted coin. But its myopic managers, like those of practically all American railroads, alienated shippers and watched in resignation as truckers captured more and more of the market for everything except commodity freight. How often did the trains run on time? How many bills that customers received were correct? How much of the locomotive fleet was in the repair shop? Uncle Pete, the railroad, didn't know. Its managers didn't keep track of those things. It wasn't part of their job. That is, until 1986. In that year Mike Walsh, the general manager of Cummins Engine -- an outsider -- became chief of this winded iron horse. Walsh, an energetic 47, is not your typical railroad guy. He was a founding member of Common Cause of California, a public-interest group. He listens to rock music in his office. When he came to Union Pacific, he kicked the company store out of its prime ground-floor space and installed a fitness center in its place. Humor and a refreshing directness save him from cockiness. What Mike Walsh has done with Uncle Pete holds lessons for any manager wrestling with an introverted corporate culture. He turned the company around, diverting managers' attention from internal lines of power and refocusing it on customer service and quality. He reorganized, eliminating layers of management and pushing responsibility down. Under pressure to meet immediate performance goals, he nevertheless made huge investments in new plant and equipment. Walsh looks to be winning. Profits have risen an average of 13% in the past two years. And customers are singing a new song. ''UP showed it's committed to modernize,'' says John Abbott, transportation manager for the FMC chemical group in Philadelphia. ''I've worked with all the major railroads, and I think UP is the best.'' THE NEW HONCHO was shooting from the hip, to an extent. Walsh admits to starting down roads whose destinations he did not know. Some decisions turned out to be wrong. But this willingness to act first and ask questions later is one of his deliberate and guiding principles. From the beginning his actions have been informed by urgency, for he knows that if he fails to push his people to the mountaintops, the parent company will invest his cash in other ventures, ones that Drew Lewis, head of the holding company, thinks will be winners, not like the railroad business, mature, old, dying.

Walsh arrived during what Union Pacific's managers thought was a very good year. Traffic was booming as the industrial economy revived, and earnings reached $385 million, a record. The executives were understandably shaken when Mike Walsh came in and told them that life was about to change. The problem: an anemic return on assets of 5.5%, which meant the railroad was not earning its real cost of capital. With an asset base of $6.9 billion in locomotives, freight cars, track, signals, and other property, much of which can be sold, the holding company had begun to invest the cash flow in trucking and hazardous waste disposal. UP's marketing, engineering, and operating departments were separate fiefdoms. The operating department, which controlled the trains, usually had the upper hand. Executive vice president Dick Davidson, head of operations, recalling those days, says: ''I and my colleagues wondered how Walsh could be so bold as to say the No. 1 objective will be to serve the customer.'' The department kept to its budget with such techniques as canceling cargo pickups and sending out trains with fewer locomotives than needed to keep on schedule. Strategic planning was an exercise conducted off in a corner solely to feed the machine at corporate headquarters. THE MACHINERY out on the line was less well attended. Locomotives usually landed in the repair shop within a month of their quarterly overhauls. Customers rejected boxcars as faulty 10% of the time. Trains derailed when new welded track expanded in the sun. The railroad accepted these mishaps as acts of God. But when the Lord sends a customer a freight car with a ventilated bottom and the customer sends it back empty and loses faith, the railroad loses money. In 1987 the cost of inefficiency and assorted screw-ups surpassed $600 million, nearly 1 1/2 times total profits. Walsh needed to gain control. His first priority: to measure how well UP performed its myriad dances -- moving trains, picking up and delivering goods, repairing locomotives, sending invoices, tracing shipments -- and to establish acceptable standards for each. He then sought to reorganize the company into a manageable form, one whose parts worked to a common goal. For help, he turned to a former head of manufacturing at General Electric. Armand Feigenbaum, now a quality control consultant, is a practitioner of Japanese-style statistical management techniques; his General Systems Co. helps companies find the demons within and exorcise them in order of their malevolence. The railroad began asking customers how it was doing. Were marketing and sales agents helpful? Did empty freight cars arrive at the right location for loading? Were bills accurate? Clients rated UP on each -- and rated its competitors too, both rail and trucking. Less than half the clients considered UP reliable. The term ''bad order,'' railroad parlance for a freight car that the customer rejects as too mangy to use, did not even exist in the trucking industry, which not coincidentally won nearly all of America's $35 billion of new shipping trade in the first seven years of this decade. With Feigenbaum's assistance, UP looked outward to see how other companies excelled. Then it surveyed their customers: those of American Express in customer service, Marriott in telephone processing, American Airlines in reservations, Hertz in vehicle preparation, Florida Power & Light in overall quality of service, Xerox in employee relations. These companies are notable not only for the service they deliver -- which competitors can see and copy -- but for the way they are managed to reach their objectives, which is not so apparent. From Marriott, for example, UP learned the best way to assess how customer reps are handling phone calls: Assign one or two people to monitor the reps all the time. NINE LAYERS of management stood between Walsh and the superintendents, the field officers responsible for the operations of trains in the territories. A super had to get permission from the bureaucracy in Omaha to spend more than a hundred bucks, and nobody outside the executive wing wanted to make a move without getting at least three signatures. Walsh eliminated six of the layers and reduced management ranks in the operating department by 800, or 50%. The superintendents now report to a single assistant vice president in Omaha, who reports to a single executive vice president, who reports to the chief. To get managers in the habit of acting quickly and innovatively, and without prodding from above, he has given superintendents power and a budget to manage; each has authority to spend as much as $25,000. This year a group of managers asked Walsh to approve a plan to forestall a shortage of grain cars. He said, ''Look, you guys are at a pretty high level in the organization. If you can't make the decision, who can? I'd rather have to rein you in than kick you in the pants to get you going.'' Walsh has sprung for some fancy hardware to help his charges do their jobs. This summer the company opened the largest computer-controlled train- dispatching facility in the world. An old freight depot in Omaha houses the $58 million Harriman Center, a concrete bunker with twin 100-yard-long black ( screens. These can display 10,000 miles of Union Pacific track in multicolored schematic form, noting the location of every train and the status of every switch on the line. Dispatchers sit in front of a section of a screen and throw switches electronically. UP built an equally up-to-date center in St. Louis for the entire customer service staff, until recently scattered among 40 regional centers. The instant a customer calls, pertinent account information automatically flashes on the service rep's monitor. Besides saving nearly $70 million a year in salary and real estate costs, the center, with its single 800 number and 24-hour service, should make life simpler for customers. They can tap into UP's mainframe to trace their own shipments and to place orders; they send in nearly a third of their own bills of lading electronically. The railroad has bought over 400 new diesel locomotives in the past two years at a cost of $500 million, the biggest equipment binge in American railroad history. It is the first company to begin equipping locomotive cabs with computers; conductors can relay up-to-the-moment status reports about cargo movements to St. Louis over a microwave and fiber-optic network, as well as receive new orders en route. Eventually the system will expand to include sensors in the track that can detect blockages on the line and reroute or stop trains. New rolling stock will also help win business from the highways. Walsh pledges that intermodal traffic -- shipments between different modes of transportation, as from a container ship onto a flatbed railroad car -- will account for one-third of UP's growth (the rest will come from growth of standard traffic). To haul the containers more efficiently, Walsh is investing heavily in so-called double-stack technology -- the simple but long-overlooked idea of stacking two layers of shipping containers for transport on rail. In June, UP completed a $15 million project, financed jointly with American President Cos., the transportation giant, and the Port of Oakland, to raise the rail tunnels east of the city for the loftier double-stack trains. Walsh's efforts have brought conspicuous changes to the way the railroad does business. It recently built a $2.7 million loading facility to serve Toyota over a short stretch in Arkansas. The business won't pay the capital costs, but the marketing department wanted to cement its ties with the automaker. ''In the old days that never would have passed the finance department,'' says Walsh. Since then, Toyota has sent more trade UP's way and is giving the railroad a chance to put together a ten-year plan for all its U.S. traffic. The railroad has begun to hit the mark in daily operations. Locomotive availability -- the percentage of the fleet that isn't up on the racks -- has improved from 86% to 93%, in effect adding about 175 locomotives to the work force. Billing accuracy has risen from 87% to 93%, though surveys show that the accuracy must reach 97% to earn a satisfactory rating from 75% of the customers. New procedures to identify danger spots on the line in hot weather and detect cracks in wheels will cut derailment expenses by 15%, or $12 million, this year. Customers are more satisfied; company surveys show they judge UP first or second against such competitors as Burlington Northern and Southern Pacific. Three years ago it ranked near the bottom of the heap. On a scale of one to ten, Boise Cascade, for example, gives UP a strong six or seven, up from three just last year. But truckers still have the edge: Boise ranks them at eight to nine. Part of the problem is inherent -- in the battle for general freight, rail will never be as flexible as highway, especially for small loads. But changes in industrial organization offer new opportunities to compete. Though truckers have been winning over devotees of just-in-time delivery by pitching their greater point-to-point speed, consistency is the most important factor for customers that are trying to link two production lines. Walsh's goal within the next three years is to match or beat truckers' reliability. THE CULTURAL changes Walsh has brought could make all the difference in meeting that goal. Handing managers more power has notably improved their frame of mind. ''We were so elated the company was willing to give us new authority that we wanted it to work,'' says Dennis Tholen, superintendent of the Cheyenne territory. ''We're not just thinking about getting the train from A to B through our territory. Now we see it as getting from A to Z.'' And as he pushes authority down, Walsh is tempering Uncle Pete's old authoritarian disposition. The new generation of superintendents sometimes sounds more Bay Area than Wild West. ''We created mistrust in our employees when we managed them through ignorance and fear,'' says Tholen. ''Employees are the only assets we have that appreciate, and we can't get it with contact management. We can do it by listening to our employees and reacting to them.'' Like other U.S. railroads, UP is trying to slim down the padded train crews left over from steam engine days. Walsh has chopped employment in the operating department from 38,000 in 1984 -- shortly after UP ingested the Missouri Pacific Railroad -- to 26,000 now. Over 3,000 of the people cut were trainmen. About 60% of the trains now run with three-person crews, vs. four, and the railroad is negotiating with the unions to increase that proportion. STILL, TRAIN CREWS are beginning to take the new corporate goals to heart. The marketing staff now routinely introduces conductors to the customers whose goods they ship. That not only leads to better service for the customer but also can add a member to the sales force at a time when consolidation has depleted UP's local presence. Earlier this year a conductor asked a customer in Shelbyville, Illinois, how he used all the paper that UP brings into his plant. When the client replied that his company makes paper cups for Burger King and sends them out by truck, the conductor said, ''Why, hell, we've got a loading ramp 100 miles away in Dupo. Why don't you truck them there and load them onto a flatcar?'' Now the client does. Setting precise performance standards has also helped get workers involved in their jobs. In the locomotive repair shop, workers were encouraged by closer relations with teams of technicians from General Motors and General Electric, UP's locomotive suppliers, who have started to work on site. With defined targets to hit and a better understanding of the equipment, the crew no longer stops as soon as a locomotive can limp out the door. The workers will now on their own initiative call up computer reports on locomotives they have overhauled to see how they perform -- just the sort of emotional buy-in that gives Walsh shivers of joy. No one can make such sweeping changes in so short a time without stumbling. In his desire to push authority down, Walsh filled the new post of superintendent with men who had received no formal management training. ''Basically, it was all hands to the wheel,'' he says. ''Because the company hadn't done a good job of decentralizing, it hadn't done a good job of bringing along younger people. I had to make big bets.'' After a year, he had to move ten of the original 30 appointees to other jobs, and he cut back the number of superintendents to 24. The phase-in of the customer service center was bumpy. Customers in, say, Idaho, who were used to calling on the local freight agent, or even yelling at him over the phone, were not at all happy about talking to a stranger on the end of an 800 number in St. Louis. And many of the service reps who moved there from around the country had never confronted the ins and outs of pricing in the official tariff, which to the uninitiated resembles the Code of Hammurabi, in Babylonian. Thus the billing errors. As is so often the case, middle- and lower-level managers have been the last to come around. ''There's a tremendous amount of skepticism,'' says a middle manager. ''We cut too deep below the superintendent level. They're constantly monitoring and changing our structure. And I've had five superintendents in five years.'' A disastrously timed change in the compensation system made matters worse. Just as Walsh was decimating managers' ranks and giving them new responsibility, more work, and higher blood pressure, he canceled all annual raises. Instead he awarded bonuses -- conditional on company performance and not part of the salary base. Walsh has decided to downplay bonuses and reintroduce raises, though some people still won't be getting one this year. Why has Walsh exposed himself to the hazards of downhill racing? He believes that if he doesn't go fast enough to risk wiping out, he loses anyway because he will miss the financial targets he has promised Drew Lewis. The yardstick Walsh selected is return on assets, partly because revenue, costs, and asset utilization are all under UP's control and partly because it's a simple measure for everyone in the corporation to understand. The choice is canny. It gives Walsh room to improve almost for free, because the railroad's assets have been used so inefficiently. He can sell off underutilized regional branches or fix locomotives more quickly or slice the average number of car-days per shipment. But there's a catch. Walsh knows that he must also make new capital investments, both to improve service and to show customers that the railroad has a future. These increase the asset base, reducing, in the short term, the crucial measure of return on assets. So, as Walsh tells his managers repeatedly, the railroad's autonomy depends on its ability to set ever higher objectives and then find a way to meet them. Walsh and Fred Henderson, the marketing director he brought in from Xerox in 1987, recently told an assembly of managers that they always expect their $ salesmen to sell a little bit more service than the rairoad can provide right now -- not to win orders under false pretenses but to set goals that will force the railroad to stretch its ability to meet customer needs, even beyond what employees think is their best effort. THUS FAR, Walsh has walked the narrow ledge between overreaching and underperforming. As profits have risen, so has return on assets -- it's now up 1.2 percentage points, to 6.7%. If the company meets its targets, earnings will continue to grow by 10% a year, and ROA will go up by as much as another point in the next three years. The railroad is making inroads into truckers' turf. A new double-stack route from Chicago to Laredo, Texas, carried 2,000 carloads a month by June, its sixth month; 15% of the traffic was stolen from the truckers. Walsh takes obvious pride that the railroad's earnings growth in 1989 will top that of Overnite Transportation, the trucker UP's holding company bought two years ago.

CHART: NOT AVAILABLE CREDIT: NO CREDIT CAPTION: INVESTOR'S SNAPSHOT UNION PACIFIC CORP.