Information technology is now the single largest capital expense for U.S. companies; overall, only labor costs more. This fact has given rise to a painful dilemma. On the one hand, companies seek to contain costs, driven by the fear that IT will increasingly devour profits. On the other hand, companies want to spend more on IT, driven by fears of internal inefficiency and competitive inferiority. This dilemma can't be resolved. The one certainty about IT, alas, is that companies will continue to lavish funds on it. Stephen Roach, chief economist at Morgan Stanley & Co., estimates that IT now consumes 41 percent of total business spending on capital equipment.
The problems that plague IT departments-- broken promises, budget overruns, project delays--aren't tolerated elsewhere. But unlike elsewhere, the world of IT is constantly in flux. Hardware becomes more powerful, software is constantly refined, staff turnover is high. Scope creep happens, as managers demand more information to run the business.
How can companies deal with such chaos? For inspiration, we present case studies of three companies--U S West Communications, Johnson & Johnson, and The Principal Financial Group-- beginning on page 43. For guidance, we offer the following seven pillars of effective IT management.
1. DON'T STARVE IT
The cost of finance as a percentage of sales is a fair benchmark of efficiency. But that's not necessarily so for the cost of IT; in fact, at any given time, the opposite may be true. "For finance, low [spending] is great," says David Axson, vice president of The Hackett Group, a Hudson, Ohio-based consultancy. "For IT, low is starving."
Bruce Stewart, vice president and research area director at IT advisory firm Gartner Group Inc., points out that British Petroleum spent significantly more on information technology in the 1980s than its peers did, rebuilding its application portfolio. Thanks to that effort, in the United States BP now spends the least on IT of the Big Seven oil companies. Similarly, Wal-Mart Stores Inc. broke an industry rule of thumb by spending in the 1980s 1.3 percent of sales on IT instead of 1 percent, but in doing so, it created the innovative distribution and inventory management systems that propelled it to the top of the retail world.
2. AVOID ROI TRAPS
You can't manage what you can't measure, and according to an exclusive CFO magazine survey of Fortune 500 companies, return on investment is a near-universal metric for measuring IT investments (see "Your Take on IT Costs," page 39). But ROI has its limitations.
Studying the ways companies measure IT investments, Bob Chatham, senior analyst at Forrester Research Inc., in Cambridge, Massachusetts, found that an overreliance on ROI leads to a predilection for cost-cutting projects over revenue-generating projects. That's because ROI measures the former with more precision, while the latter involve forecasting and uncertainty.
But certainty may have a big opportunity cost. Writes Chatham in a Forrester report: "For the typical company, there is often a factor of 10 times greater profit leverage in revenue- generating applications that enable companies to add value and raise prices than in those designed to cut costs."
Meanwhile, growing awareness of the total cost of ownership (TCO) of computers has made ROI calculations both more accurate and more difficult. "We used to have trouble measuring the return," says Jerry Kanter, director of Babson College's Center for Information Management Studies (CIMS), in Wellesley, Massachusetts. "Now, we [also] have trouble measuring the investment." Distributed computing results in distributed costs, including some that CFOs might not be aware of- -in particular, maintenance, training, and support.
3. DON'T ASSUME TECHNOLOGY IS THE ANSWER
Reports of spectacular ROIs almost always involve inventory management systems, notes Mark W. Doll, New Englandarea director in the management consulting group of Ernst & Young LLP. That's because the ROI is easy to track. If a $2 million inventory system allows a company to identify $100 million in stock that it doesn't need, and if the company subtracts the carrying costs on that stock over five years--"Bang! You've got a high ROI," says Doll.
But a high ROI from a new inventory management system can also be seen as an indication that a company is a poor manager of working capital. Why not adopt sound management practices before leaping to an expensive technology solution?
"Technology is often foolishly viewed in isolation," says Axson. Any application can promise an attractive ROI, "but you have to ask, is this a problem that technology should be solving?" For example, if 16 signatures are required to approve a capital expenditure, workflow software can automate that process. "But why do you need 16 signatures? Because the 16th person doesn't trust the other 15," says Axson. "Simplify that process first, then add workflow if needed."
4. MEASURE, MEASURE, MEASURE
It's one thing to say you will reap benefits from an IT project; it's another to actually reap them. "Most enterprises do a lousy job of actually getting the benefits," claims Gartner's Stewart. There are three main reasons why this is so.
First, the business case doesn't specify the metrics--financial, operational, or process-- that will be used to evaluate the project. If a project is supposed to reduce head count, says Stewart, the affected employees should be named. If a sales force automation system is expected to make the sales force more effective, you should see sales volume and repeat business increase, as well as average order size and even "face time."
Second, especially in the case of cross- functional applications, the changes wrought on people's jobs by new technology aren't anticipated. If work isn't reengineered in advance of, say, an SAP R/3 installation, organizations may actually take a productivity hit, says Stewart. It's not enough to justify a new application on the basis that employees will gain more time to do value-added work. The questions are, exactly what value-added work will they do, and how will you capture the benefit of that?
The third main reason IT proj-ects don't deliver benefits is that business sponsors don't follow through. A humble yet effective tool in this regard is the user survey, says Kanter of CIMS. Ask key users and project champions: What gains were there in savings or revenues? Did services increase? What problems did users encounter, and how can they be ironed out?
5. REMEMBER THE INFRASTRUCTURE
With all the attention given to the risk and reward of large projects involving glamorous new technology, it's easy to forget where the rest of the IT budget goes. On average, 60 percent of a large company's IT spending is devoted to the operational infrastructure, says Alan G. Gonchar, president of Compass America Inc., in Reston, Virginia. That includes data centers, wide area and local area networks, the desktop, application development and maintenance, and labor.
To keep infrastructure costs in line, efficient IT organizations employ a number of tools:
- benchmarking, to determine the true costs of IT and establish a baseline for improving operations;
- technology standards, to ensure consistency of platforms, software, and spending;
- centralized procurement, to maintain control over purchases and manage suppliers;
- asset management software and services, to track and depreciate hardware and monitor software licenses;
- outsourcing, to farm out functions that external vendors can handle more efficiently.
A goal of many of these practices is to reduce complexity, the unacknowledged legislator of infrastructure costs. Companies often take a fragmented approach to IT, observes Axson, allowing business units to build technology empires--choosing their own hardware and customizing software to their particular needs. In so doing, they risk forgoing opportunities to reduce development, maintenance, and training costs, and to share data and best practices.
When benchmarking infrastructure operations, be leery of "high-level" benchmarks, says Gonchar. Unless IT is analyzed on a unit-cost basis--such as the cost to produce 10,000 pages of laser output, or the cost per help- desk call--"you'll never figure out how to improve."
6. RETHINK CHARGEBACKS
Ideally, chargeback systems help users understand their IT spending and help managers allocate resources efficiently. But they often fail on both counts. According to a 1996 Forrester Research study, 85 percent of Fortune 1,000 companies said that chargebacks ranged from "not effective" to "moderately effective."
One reason IT chargebacks fail is that they take the form of flat allocations, spread evenly across the user population. This approach merely gives departments incentives to increase usage, points out Steve Player, firmwide director of cost management for Arthur Andersen LLP, in Dallas. As individual users attempt to "get their fair share" of an arbitrary percentage, whether of revenues or machine time, total IT usage goes up, creating a vicious feedback loop.
Effective chargebacks are assessed only for services that employees actually use, and in terms that allow them to identify opportunities for savings. For example, says Player, instead of charging users for CPU seconds, charge them on a per-report basis. Or instead of assessing one total cost per PC, break that cost into its component costs.
A chargeback is "a tool to increase communication," says Player. But communication occurs only if everyone speaks the same language.
7. MAKE THE CIO YOUR FRIEND
In many organizations, there is friction between IT and finance. CIOs are stereotypically perceived as always asking for money, to be spent on some new technology in search of a solution. Yet CFOs and CIOs are natural allies; finance, like IT, is one of the few functions that cross all corporate boundaries. CFOs and CIOs should be working together to promote sound IT management practices and to ensure that IT supports strategic aims.
The CFO shouldn't let the CIO become the fall guy for failed proj-ects. The business users of technology should ultimately be responsible for the success of IT projects. The CIO should be responsible for a well-run IT department that delivers realistic technology solutions on time and on budget. And the CFO should understand where and how IT can add value--and how to measure that value.
Indeed, several former finance executives have understood the value of IT well enough to become CIOs themselves. David R. Laube, CIO of U S West Communications, has a background in finance. So do other prominent CIOs, like Richard J. Fishburn of Digital Equipment Corp., Rick Bentz of Hershey Foods Corp., and InformationWeek's current CIO of the Year, Denis O'Leary of Chase Manhattan Bank. Laube endorses the switch.
"The CFO understands the business and knows where the key leverage points are to drive the business to success," he says. "If the CFO has a strong understanding of technology, I think it's a natural extension to move to the CIO's job."
This is not to suggest that companies ought to select their CIOs from the ranks of finance. But as IT becomes an ever-larger line item, CIOs need a stronger grasp of finance, and CFOs need a deeper understanding of how technology supports the business.
YOUR TAKE ON IT COSTS
In May, CFO magazine sent a survey on information-technology spending to the companies on the Fortune 500 list. Forty-nine companies responded.
These companies will spend an average of 2.2 percent of annual revenues on IT in 1997, ranging from a low of one-half of 1 percent to a high of 10 percent. More than a third of the companies, 37 percent, will spend no more than 1 percent of revenues on IT.
In 1998, the respondents expect IT spending to rise by an average of 8.3 percent. Here, averages are less telling than distribution. A healthy 40 percent project IT spending to stay the same or fall in 1998.
What are companies spending their money on? Eight companies named infrastructure improvements as their most expensive current project. Six named financial systems. Five specified SAP implementations as their most expensive current IT project, costing from $25 million to $120 million. A sixth project, the biggest in the survey, involves rolling out SAP with other applications. Cost: $200 million. Four companies are putting their money on customer-related systems. Three specified year 2000 conversions, at a cost ranging from $3 million to $25 million. The cheapest current project in the survey will cost $300,000.
Overall, the sheer size of these projects indicates just how much Corporate America is staking on IT. Eight companies are engaged in projects ranging in cost from $50 million to $75 million; five companies are in projects from $100 million to $150 million; and one company, cited above, is in a $200 million project.
How do these companies justify their spending? Nearly all companies use ROI to evaluate major projects; only 4 said they didn't use ROI at all. ("How do you calculate the return?" asked one naysayer.) But only 16 companies (33 percent) said ROI was one of "the most important measures" they rely on to evaluate IT expenditures. (As for other financial metrics, 4 companies cited payback periods, 3 cited IRR, 2 named EVA, and 1 specified NPV.)
Respondents listed several other important measures, ranging from productivity improvement to cost cutting to customer- service enhancement. Twelve companies (25 percent) cited strategic support as an important measure of IT projects. But our favorite response was submitted by a diversified manufacturer as the single most important measure of IT: "Is it simple and needed, or is it driven by an overly complex business structure?"
Finally, we asked companies to estimate what percentages of IT projects were primarily aimed at cost savings, revenue growth, or "soft" benefits (for example, increased information or improved quality). Their response: on average, 50 percent of projects are primarily aimed at cost savings, 28 percent at revenue growth, and 24 percent are aimed at soft benefits. (These numbers are rounded; some firms' responses exceeded 100 percent.)
US West Communications
"When you try to measure things, in kind of a bean-counter way, you're missing the point," declares David R. Laube. "The point is, how do you use information technology to drive strategic advantage? And how do you know if your IT function is performing at a high enough level to make that happen?"
Those are the questions Laube tries to answer at U S West Communications, the Denver-based, $10.5 billion (in revenues) regional Bell operating unit of U S West Inc. Not that the 49-year-old vice president and CIO is unsympathetic to a finance executive's concerns: Laube previously served as vice president and controller of Pacific Northwest Bell and as CFO of Mountain Bell, both of which were merged into U S West in 1983. And he had been vice president, controller, and treasurer of U S West Communications for five years when he was "drafted," in his words, to become CIO.
"IT wasn't delivering what it needed to deliver for the business," recalls Laube. "We needed a business-oriented person with a strong understanding of technology. I had quite a bit of systems experience." The move paid off: between 1995 and 1997, Laube helped reduce annual spending on the IT infra- structure by $300 million.
Laube's first priority as CIO was to impose "business disciplines" on the IT function. U S West would develop and conform to technology standards (immediately reducing 14 E-mail systems to 2, for instance), use its size to leverage software and hardware suppliers, and avoid bleeding-edge technologies that hadn't been proved in the marketplace. The IT function would become outward-focused and client-oriented, explaining technical risk in business terms and adhering rigorously to budgets and schedules.
Above all, Laube wanted to ensure that U S West would spend the "right amount on the right things." Project sponsors must build business cases for their proposals, which compete with one another during an annual resource allocation process. But U S West doesn't simply rank proposals and dole out the money until it runs out. Instead, it "optimizes" projects.
"We first come up with a utility value based on the business case," explains Laube. A "modified profitability index" is used to rank proj-ects financially, based on expected cash inflows divided by cash outflows on a discounted basis. ("ROI is just not relevant," he maintains. "It's not a very good measure-- what's the R, and what's the I?") Subjective benefits, such as a proj-ect's contribution to quality or to business success, are converted into numbers, and all measures are run through an algorithm to arrive at a project's utility value.
"But the most highly rated project may consume all your resources," notes Laube. Like a professional sports team that trades a top draft pick for a number of lower, less expensive ones, U S West might pass over a project with the highest individual utility value in favor of several lower-rated proj- ects with a superior combined utility value. Hence, "optimization."
The benefits claimed by the business units for the approved IT projects are put into targets for the year. "In effect, you automatically track [the projects]," says Laube. "The business units need to be the advocates for IT spending," he adds. If the units can generate enough value [from a project] and sponsorship for that value, then they are going to be the ones that are stepping up to the commitments; they are putting the benefits in the plan."
A multiyear IT project is particularly challenging, observes Laube, because the budget process is annual. "The good part is that it requires the project to continue to stand up to financial scrutiny," he says. The bad part? "It's a distraction to the people involved. Nobody wants to think that [the project] could potentially not be completed. Motivation and productivity become issues."
Some multiyear projects are so important that they don't have to be rejustified. Currently, the single largest such project at U S West is the consolidation of its three legacy telephone billing systems--a three-year project that began in 1996. Cost: in excess of $100 million.
What does Laube expect the CFO to know about IT management? He prefaces his reply by affirming that both he and CFO Al Spies are on the same wavelength (both report to CEO Sol Trujillo). "I expect the CFO to understand the linkage between business performance and the demand on IT--the demand for deliverables. Does he have confidence that IT can deliver? Does he have confidence that the business will be able to generate the benefits involved? If the CFO measures IT just on how much it's reducing its spending, that could absolutely be the wrong measure."
DAVID LAUBE'S TOP 10 WAYS TO CUT IT COSTS
- Obtain control of all IT resources.
- Consolidate all systems procurement.
- Develop project-based control of IT spending.
- Link all IT projects to business value.
- Use technology that works.
- Develop and enforce technology standards.
- Aggressively consolidate computer operations.
- Outsource, but only where it is appropriate.
- Ruthlessly prune redundant and overlapping systems.
- Deploy strong project-management methodologies.
Source: U S West Communications
THE VALUE CHAIN
Johnson & Johnson
Every dollar spent on information technology should achieve at least one of three goals, according to Edward B. Parrish: reduce costs, increase information, or improve quality or productivity. Sounds simple, but that's an extremely challenging goal for Parrish, corporate vice president of information management at Johnson & Johnson, the $21.6 billion health-care-products giant based in New Brunswick, New Jersey.Until two years ago, Parrish was vice president of information architecture, not management. Why the change in title? "We're trying to emphasize the utilization of information technology in business," says Parrish. "We want to change the focus from providing technological functionality to providing solutions to business problems."
The concept or "paradigm" that is helping Parrish to change the IT culture at J&J is the value chain. The chain consists of four major links: in ascending order, they are infrastructure, transaction processing, information management, and revenue growth. More than half of J&J's IT expenditures is devoted to the first two links, which Parrish calls efficiency areas. "But they return the least business value," he says. "The challenge is, how do you lower the amount of resources in the first two to support the second two [which are considered investment areas]?"
Johnson & Johnson is a highly decentralized organization; its 172 operating units create their own IT budgets and develop business cases for IT investments. This year, for the first time, the units are structuring their budgets according to the four areas of the value chain.
"Once you've created this paradigm, you can sit down and discuss [IT spending] as businesspeople," says Parrish. "Formerly, management at corporate would look at the budgets according to what percentage [of sales] was being spent on IT. Two percent might have been okay; rising percentages were not okay." Now, IT spending is no longer viewed as monolithic. "If you're spending 2 percent on the efficiency levels and less on the top two levels, that's not necessarily good."
For applications in upper links of the value chain--information management and revenue growth --Parrish wants the operating companies to think in terms of investment decisions, as if they were starting a new product line or launching an ad campaign. Accordingly, the metrics used to monitor those investments are business-oriented and output-based: cycle time, time to market, information availability, and so on. Working out the metrics has been a joint project of information management and finance, first under former CFO Clark Johnson and since last February headed by Johnson's successor, Robert Darretta.
This is not to say that Parrish disregards aggregate measures, like IT costs as a percentage of sales. "I track them and keep them in my pocket, but I don't like to talk about them," he says. Parrish believes that if you start worrying about IT costs in the aggregate, you'll try to "manage" them without regard to specific business cases. "I want the discussion to be on the business side," he says.
CLASSIC COST CONTROL
The infrastructure (networks, telecommunications, data centers, PCs) is subject to "classic cost control, like that in a manufacturing plant," says Parrish. Similarly, the aim in transaction processing is to lower the cost per transaction. An executive committee watches a comprehensive set of unit-cost metrics that measure the performance of mainframe computing, global networking, and so on. The committee regularly reviews cost trends and compares them with benchmarks provided by external benchmarking firms.
"We encourage some sharing and leveraging of infrastructure planning and standards among companies," says Parrish. For example, technology standards for PCs are set and promulgated by Parrish's office. The operating companies take those standards and work with purchasing to negotiate the lowest possible contract from vendors. "We give companies the freedom to commoditize the product," says Parrish.
If J&J can't provide an IT function cheaply enough, it will outsource it, says Parrish. Outsourcing, he warns, shouldn't be considered a tonic for poor management. "If you don't know how to manage [a process] yourself--if that's the reason you're handing it to an outsourcer--that's a bird's nest on the ground." Translation: the outsourcer, not the company, reaps the savings from the cleaned-up process.
If Parrish could reduce his management approach to an algorithm, what would it be? "First, divide up the spending. Divide up the objectives. Look at the distribution of your spending according to the four components in the value chain: infrastructure, transaction processing, information management, and revenue growth. Sometimes you can fool yourself. Drive the lower two links in the value chain for efficiency, as hard as you can. The other two: push for effectiveness."
The Principal Financial Group
First principle: The Principal Financial Group believes in partnerships between businesspeople and techies. "We don't go forcing things on business units," says CFO Michael H. Gersie. "We just installed a data warehouse in our group business unit, for analyzing claim data. It was installed with the full participation of businesspeople and financial people on the group side. It's our job to help deliver that capability."If it doesn't sound clear which side Gersie's on-- finance or IT--that's because he was The Principal's CIO until last July, when he became the first CFO of the Des Moinesbased, $19 billion mutual insurance and financial services company. Even now he's a walking partnership, wearing two hats until a new CIO is named.
Gersie, whose 27 years with the insurer include stints in finance, tax, and new- business development, was named CIO in 1994. "They were looking for someone with a broad business background who knew something about technology," he recalls.
Gersie was also charged with developing corporatewide IT standards. "The way we developed standards is similar to the way that countries organize," he says. Infrastructure matters--the network and data center that link the organization--are handled at the "federal," or corporate, level. Standards for application development and hardware and software purchases are maintained by the "states," or business units, which have their own CIOs. "Each business unit is in charge of its own fate," says Gersie. "Ultimately, [the choice of standards] is going to affect that unit's bottom line."
NO HURDLE RATES
The Principal watches its IT spending carefully, but "we've always recognized technology as playing a key strategic role," says Gersie. "We've never been starved for technology dollars. We would be extremely shortsighted to clamp down on our IT budget. That's been one of our strengths."
The CFO doesn't enforce a corporatewide methodology for evaluating IT expenditures. Big projects are usually analyzed in terms of discounted cash flow and return on investment. But is there a rigorous hurdle rate? "No," he says. "We recognize that these can be very soft investments. We try to do a good job of estimating costs, but it's difficult to get real close to exact costs." For a major project with a one- or two-year timetable, "you're making some assumptions about the leap of productivity you're going to get," says Gersie. "You can cost it out, but your information is only as good as your assumptions."
Currently, the all-in cost of IT at The Principal is around $180 million, or 1 percent of annual premium revenues. The growth rate, which was in double digits, has slowed down, but it's still in the high single digits. "Like everyone else, we have had trouble getting the staff we need," Gersie notes.
Gersie is proud of the progress the insurer has made on what is currently its biggest IT project. Two years in the making, it's an imaging and workflow application that will be integrated with The Principal's pension administration system. As a full-service 401 (k) administrator, The Principal is on the receiving end of a massive paper flow generated by plan participants. The new application, dubbed Pension Express, is designed to process within 24 hours every piece of paper received.
"The pension area sees it as a strategic application," says Gersie, providing better, more responsive customer service. As such, it was justified primarily in terms of revenue generation and service quality; no official ROI was calculated.
An internal lab was set up to do extensive modeling and prototyping. Then, like a Broadway producer, the insurer took Pension Express on the road, trying it out in remote locations. In June, it began installation at headquarters in Des Moines. When The Principal completes the first upgrade by September, the application will run on about 2,000 workstations. Estimated total cost: around $30 million.