Wang notes that successful programmers develop not just good software but big egos. A hit product can send its developer off, wanting to use his or her clout to explore new directions, leaving the lieutenants to write the next -- invariably inferior -- release of that product. Wang claims that the initial creator of the program can best understand the logic and inspiration behind it. So he or she should write the follow-on product.
Wang pays his star programmers as much as $250,000 to reward them for good work and to underscore that they don't have to move to management positions to make more money. He also limits the scope of each project. "I never let my programmers do the whole thing in one fell swoop. I cut it back," he says. "I leave a carrot out there in the form of the second release. Nirvana is not achieved in one stroke."
To further streamline development, Computer Associates has spent more than 15 years devising a core architecture called CA 90s, around which it can standardize the products it writes. Such a setup not only creates uniformity across the product line but also precludes programmers' having to start from scratch with each development project. Russ Artzt says CA 90s has imposed a discipline on Computer Associates, something not often seen in software companies, where programmers so readily fall in love with their creations.
CA 90s contains what Artzt labels "reusable" components, a concept that applies in a different sense at Nucor. Nucor continually upgrades its existing equipment, instead of making wholesale technical transformations. Sitting over lunch in the cafeteria alongside coworkers Frank Pugh and Richard Painter, Glenn Pushis says, "You're looking at three-quarters of the engineering department." Their job is to persistently troubleshoot at the plant and spot equipment upgrades they can make that will pay for themselves in savings within a reasonable time frame. Most proposals costing less than $5 million get immediate approval.
The Nucor engineers have customized practically the whole second casting line at the plant, based on their observations of how the first one has run. They have been able to up the amount of steel moving down the line from 220 inches a minute to 250. The engineers, in concert with production workers, keep the entire plant in perpetual upgrade. Nine months before the galvanizing line started up, Nucor assigned production workers to consult with the engineers on its design and construction.
"We look at every piece of equipment and make it go faster or increase its reliability," says Kevin Young, the cold-mill manager. Young, who assumed that job last October, says that since then his department has increased productivity on all six of its lines. Output on the pickling line is up 10% since October. The reversing mill is up 10%. The tempering mill has increased production by 25% -- with 25% fewer workers.
Valera Shifrin, one of Nucor's three metallurgists, earlier this year began rethinking the company's deoxidation process, which removes impurities in molten steel. Ultimately, Shifrin's idea ended up saving 50¢ per ton of steel. That sounds small, until you consider that Crawfordsville will produce 1.8 million tons of steel this year -- and will save nearly $1 million.
Results like those achieved by Shifrin -- an unlikely recruit who was a citizen of the Soviet Union when he first read about Nucor in an academic journal -- are hard to argue with. But not impossible.
If there is a downside to bootstrapping in big companies like Nucor and Computer Associates, it is that every move by management becomes a return-on-investment issue and little else. Asked about the apparent incongruity of the comfortable layout at Computer Associates with the conventional bootstrapping image, Charles Wang immediately responds, "We have the highest revenue per employee in the software industry. That says we are efficient."
Wang goes on to decry companies at which a so-called roach coach pulls up in the parking lot outside and employees stream out of the building to buy a cup of coffee or a sandwich. They are wasting time. He says that offering employees a range of amenities gives them a predictable environment that allows them to focus on work. "We try to provide everything for everybody," he says.
The feeling that companies like Nucor and Computer Associates often project is that the organization always knows best. Philip Montrowe abruptly left Computer Associates six years ago after a meteoric rise at the company "because I got mad at Charles. I didn't want to be reinvented at that time." The company wanted to move Montrowe, whose expertise was in mainframe-computer-related software, into another area. Five years later he returned, conceding that Wang had been right. But Montrowe labels Computer Associates a place "where I guess you can more or less speak your mind." Though he calls Wang "brilliant" and "visionary," he also adds, "you wouldn't want to get on his bad side."
The same can be said of Nucor's management. Sitting over coffee one day at Crawfordsville, cold-mill supervisor Paul Rokosz details how one employee, who was suspected of drug use, took a urine test, which turned up negative. He was later confronted with the prospect of a hair-follicle test. "At that point he resigned," Rokosz recalls, adding, "we screen our employees heavily."
The message at these companies is that slackers need not apply. They are the Marine Corps of corporate America. And yet Nucor and Computer Associates simply mirror the rigors and perils of an economy in which bootstrapping and downsizing are beginning to converge. With companies asking their workers to do more with less, the effort to keep people motivated "will become the human-resources issue of the '90s," predicts consultant Susan Rowland.
Rowland adds that for all the corporate downsizing and cost slashing that has occurred, the bootstrapping phenomenon "has really only just begun." But at Nucor and Computer Associates, it never ended.
