Making Enterprise Software Work
CEOs have learned many hard lessons about installing software.
July 1 2004 by Chief Executive
What’s worse than going to the dentist? Having to do the drilling yourself. For CEOs and other corporate leaders, the equivalent of that is implementing enterprise software. So what Harold Karp, president of Essex Electrical Products, faced had to feel like a self-inflicted root canal. Sold in December 2002 by Superior TeleCom to a holding company called The Alpine Group, Essex, one of the country’s largest cable and wire manufacturers, left behind all of its legacy information technology systems in the transition. There was no choice but to go through “the pain and agony,” as Karp puts it, of switching to new enterprise systems.
But Karp had suffered through previous information technology misadventures, and prior agony informed his approach. He knew that the only way to keep things on track after installing a new enterprise resource planning package from QAD and pricing software from Metreo was through active supervision from the executive suite. In the past, Karp had tried delegating such oversight. “I’ve had teams that were headed up by just a business executive and I’ve had them headed up by just an IT executive,” he says. In both cases, he adds, the teams achieved only “limited success.”
This time, he handpicked a five-person team of top employees€¦quot;not managers€¦quot;from various business functions and devoted them to the project. One of the five led the project and reported directly to Karp, who acted as the executive sponsor and met with the team weekly to make sure they were getting the necessary support and making progress. Ultimately, it worked. “We designed it to get the data we needed to make the day-to-day and week-to-week decisions,” says Karp, whose private company is based in Fort Wayne, Ind., and has annual revenues of $300 million. As a result, IT operating expenses dropped by 35 percent while sales gained 1 percent, which translates into margins of 2.5 to 3 percent.
|DOs & DON’Ts|
SOURCE: Chief Executive
For most companies, implementing enterprise software doesn’t work so well. Gartner reports that as much as 20 percent of the money that businesses spend globally on technology is a total waste. The archetypal disappointment was Nike’s failed implementation of an extensive supply chain system in 2000, which led to CEO Phil Knight’s now famous lamentation: “Is this what we get for our $400 million?”
The difference between success and failure isn’t only a matter of financial commitment, or even the technical prowess of a corporation. What really makes the difference is the CEO. It appears that getting a return on investment from software has more to do with well-defined business goals and strong project sponsorship from the top than it does with the latest technology. Successful software deployments also need the best cross-functional talent and resources. According to a Cap Gemini Ernst & Young polling of hundreds of executives in late 2003, some 60 percent said they either have or were planning to have a business executive run IT so that it is less of a “money pit.”
There was a time when many of the problems with enterprise software were largely the responsibility of the vendor. “I think they oversold what was standard in the package,” says Greg Owens, chairman and CEO of Manugistics, a $243 million supply-chain software company in Rockville, Md. But Owens says that rampant overpromising and underdelivering by software companies seem to have largely subsided after the shakeouts of the ’90s.
Now, some analysts say, companies have no one to blame but themselves when implementations fail. One of the problems is that despite all the failed whizbang projects, too many purchases are still made on a whim. Laura Preslan of AMR Research talks of “management by airline magazine,” referring to how a CEO might read an article during a flight and, upon returning to the office, march into the CIO’s office demanding the latest, greatest thing.
Companies are still jumping in without thinking through the implications, agrees Michael Pugliese, managing director of the PeopleSoft practice at the consulting firm BearingPoint. “To just say we’re doing CRM is not a meaningful event,” he says. Any major software package has a range of potential uses. Customer resource management can help increase customer satisfaction, manage buyer lifetime value and provide information to guide marketing tactics. ERP packages offer expanded data availability, potential operational efficiency and cross-functional integration. Supply chain software can increase availability of product, decrease inventory costs or speed the time between an order and fulfillment. Here are some lessons that CEOs have learned.
Keep a tight focus. Too often, companies try to emphasize everything and wind up with too many priorities. Business strategy must dictate the proper emphasis. When the organizational imperative is overly simple€¦quot;”Do CRM,” “Do data mining,” “Do supply chain” or “Do ERP”€¦quot;then lower levels of the company are left to interpret the directive and implement it. Senior management must be clear and precise about what the new system is expected to accomplish.
Similarly, in large corporations, different business units have their own needs and interpretations of the company’s direction. “That’s why you end up with so many differentiations of ERP within the same company, even though a single instance would have been easier to support and provide the base level of functionality that a company wants,” says Manugistics’ Owens. The result, according to Robert Zahler, a Washington, D.C.-based partner at the law firm Shaw Pittman who has seen many of his clients buy enterprise packages, is often an underestimation (“by 50 to 100 percent”) of the eventual price tag. Because the cost can run from tens to hundreds of millions of dollars, the impact on the bottom line is enormous.
CEO involvement is key. Karp achieved a kind of pyramid model at Essex Electrical Products during implementation, in which the influence of his vision spread throughout the organization while feedback from below bubbled up. This kind of structure has become vital to the way businesses must make IT decisions. “When we all grew up in consulting and business, you defined your strategy and you then articulated the actions and technology and processes to facilitate that,” says Christopher Formant, a BearingPoint executive vice president. Now, increasingly, CEOs set the strategy, influenced by the possibilities of technology. As software moves into place, further feedback can redirect their thinking about strategic timelines and tactical goals.
Extending the pyramid into software’s actual use is also imperative, but tricky. Most legacy software packages are actually designed for particular departments, not for the CEO. “I might only want to see a couple of key metrics,” says Jeff Ferguson, president of management and operations for Erickson Retirement Communities. “If the implementation is left within a department and there has been little if any involvement with the CEO, then [the CEOs] spend their time hunting and pecking to find the information.” Critical data might also be locked away on desktops in spreadsheets and never find its way back into the application.
Lack of two-way communication also threatens operations. Management designs and implements processes to reach business objectives. Any large software package is intrinsically linked to those daily activities. In the most positive scenario, a company learns best industry and functional practices from the software and improves. At worst, the enterprise package can potentially uproot current working processes. Without communication, upper management will find out too late what is actually happening.
Keep it realistic. Ambition can overreach abilities either at the start of a project or else through function creep, where people keep adding things on their “nice to have” lists. When a company bites off more than it can chew, it can choke. “If you try to do too much, you’ll never get the app in,” says Max Hopper, a board member of Fuego, a software provider in Plano, Tex. He’s also the former chairman of the Sabre Group and US Data and the former CIO of American Airlines and Bank of America.
With implementation timelines running a year to 18 months or even more, not everything will get done at once nor does it have to. Smart management creates priorities and focuses on areas that can provide the quickest and largest return on investment for the corporation as a whole, and not for a particular department. Some CEOs, like Yellow Roadway’s Bill Zollars, insist on neatly defined IT projects that last no longer than a year. “I’ve been involved with way too many three-year projects that turned into five-year projects that didn’t return the investment they were supposed to,” Zollars says.
Customized or off the shelf? Another major decision is whether to customize a vendor’s software. A company that resists changing its own processes may be tempted to modify the software, but there are hazards to that approach. In the long run, that company gives up something important: the ability to apply new releases of the package without rewriting them as well. “Go through a thought process that asks, €˜Is this so valuable to me that I will customize it?’” says BearingPoint’s Pugliese. As much as possible, companies are better off with generic implementations. Does it ever make sense to customize? Certainly, if there is some core function that’s critical to a company’s operations.
But whether large or small, most companies are better off using applications as others have designed them. As software categories have matured over the years, more can fulfill customer needs out of the box. Companies should focus on filling 75 to 80 percent of their needs from an off-the-shelf package.
Anticipating trouble. While you can’t foresee every crisis, it’s a good idea to keep an eagle eye for holes in your strategy. When American Airlines was switching from single-application mainframes and dumb terminals to PC-based user platforms that would run multiple packages, the effort was blindsided. “We stubbed our toes forgetting all the elements we should have thought of,” Hopper recalls.
For example, the company rolled out the technology on a floor-by-floor basis, even when a particular business unit was split between different floors, “which is kind of stupid when you think about it,” he says. The IT group had to constantly switch the hardware, applications and training they were providing, rather than focus on getting one group of related users squared away. Eventually, the company switched to rolling out the systems by working groups, regardless of where the employees were situated.
Handling the people factor. There are, of course, situations that even experience cannot circumvent€¦quot;for one, employees reluctant to transition from the current system. Harrah’s Entertainment of Las Vegas has been the poster child of data mining applications, increasing revenues and customer retention, using specialized systems from Teradata that regularly analyze tens of millions of database records. For CEO Gary Loveman, a former academic with an expertise in statistics, that was a natural fit. But many employees either could not understand the new techniques or resented change and the end of marketing fiefdoms at the company’s different properties. “In the corporate marketing department, I don’t recall that there were more than one or two survivors,” Loveman says. Early on, the field marketing groups lost a third of their staffing and continued to require replacements over time. In other words, sometimes the people have to change€¦quot;or be changed.
When applications clash. In many industries, enterprise systems will not work together, and that can take a toll. Ferguson of Erickson Retirement has found that the financial reporting software from Hyperion does not link with sales and marketing information from CRM software or with data from clinical systems. As a result, Ferguson isn’t able to review the data for all 10 retirement-living campuses in a timely manner. “It’s usually the middle of the month before I see all these reports at the same time, so there’s a time lag,” he says. Ferguson estimates that even getting the data a week and a half earlier could translate into a 20 to 30 percent improvement in decision making, and complete integration could cut by two days his current review time of almost a week every month.
The problem is that it takes time to extract the results from each of the different software systems. Without having integrated systems, there is really no way for a company like Erickson to break out of this bind. In what is almost certainly a classic understatement, Jay Coughlan, president and CEO of Lawson Software, an enterprise management software company based in St. Paul, Minn., says that “companies do not consider major enterprise software systems to be very easy to integrate with their existing systems.” He acknowledges that enterprise software companies still encounter problems when tying systems together. Although the situation is better in some industries than others, Coughlan says, there’s still room for improvement.
There may be no ultimate answer to the question of who’s fault is it that enterprise software is so difficult? At least some of the responsibility may circle back to the purchasing company if it is not sophisticated in dealing with the impact of technology. Fewer than a third of the companies that Lawson has talked with even did the baseline performance measurements to see whether installed software actually offered a return on investment. And based on the information that comes from newly installed systems, only a handful go through the changes necessary to improve their operations.
Installing major enterprise software often isn’t easy, but the potential benefits can be enormous. Proper involvement of and oversight from a CEO can help ensure that the payoff outweighs the pain.