CFOs to tech: 'I'll spend for the right technology'.
In the 1970s, they spent big on minicomputers that freed companies from the tyranny of a centralized data-processing department. In the '80s, PCs led the way by unleashing a torrent of investment to increase knowledge-worker productivity. And, in the '90s, the Internet connected those PCs into a global network that contributed to trillions of dollars of investment and millions of new jobs
Each of these then new technologies spurred business--and more recently, consumers--to invest, due to the boost to their productivity. But this decade-long rhythm of IT innovation has been interrupted. These days, companies haven't found a new wave, and analyze rigorously before investing in technology. But once the money is spent, they don't rigorously track actual returns to see if those matched projections.
Dewey Norton, director of finance at San Diego, Calif.-based Instant Information Systems (IIS), suggests that instead of tracking its investment, IIS asks whether a system lets it make better decisions. According to Norton, "A system may not have a quantitative impact. We'll ask the business people, 'What did you accomplish with this afterward? Did you have the information you needed? Was the information right? Was it timely?'"
Quantifying an IT investment payoff is not so clear-cut. Meta Group's 1,700-company database calculates that revenue per employee and income per employee have increased 48 percent (from $324,000 to $478,000) and 14 percent (from $22,000 to $25,000), respectively, since 1997. Meta Group's Executive Vice President Howard Rubin and Battery Venture's General Partner Dave Tabors suggest that IT accounts for most of this productivity improvement. But neither can prove it. (Rubin's estimates, in the chart on the next page, suggest that companies will spend enormous sums on IT this decade, whether or not some profound technologies emerge.)
Curiously, companies do spend on technology that promises to cut their IT budgets! For example, Kavin Moody, director of Babson College's Center for Information Management Studies (CIMS), suggests that companies have recently reduced some of their IT costs by a factor of 10 by replacing expensive servers running proprietary software with cheaper blade servers running on Linux.
Also, Rubin suggests that since 2001, cost-conscious companies have gone through the following waves of IT cost-cutting:
1. They cut IT budgets by reducing IT staff;
2. They squeeze the existing IT infrastructure by investing in technologies that cut IT budgets, such as the one Moody mentioned;
3. They outsource IT;
4. They offshore IT;
5. They bring IT back in-house and "run it right."
Today's IT finance executives echo this decade's emphasis on tight control of IT budgets. The director of IT finance for a $90 billion global company is proud that his company's IT budget is a mere 2 percent of revenues, about half the 4 percent to 5 percent average. This company budgets two kinds of IT projects: "lights-on," which are necessary to keep the business going, and "development and enhancement." Before funds are available for the latter, all lights-on projects must be funded. The company allocates its development and enhancement budget based on a value rating consisting of three key tests:
1. ALIGNMENT WITH STRATEGIC INTENT: How a project contributes to revenue growth or cost reduction goals;
2. RISK AND COST: Whether a technology is poorly supported or backed by a financially weak vendor; and
3. QUALITY OF SERVICE: How a project affects system availability, for example, at a time of the month when demand is particularly high.
Despite intensive scrutiny of its IT budget, the company does not go back and conduct post-completion audits. As a result, if a project plan predicted, say, a 150-person headcount reduction, the company would have no way of knowing whether the planned outcome had, in fact, been achieved. The executive says that these planned headcount reductions tend to "lose their identity." Although it can track how many transactions its systems process, it can't tell for sure whether IT projects' business benefits have been achieved.
Does the company see any exciting IT developments on the horizon? Despite receiving many sales pitches from IT vendors, the short answer is "not much." This moribund attitude is echoed by Lorin Chevalier, CFO of American Specialty Health, a privately held $100 million provider of complementary healthcare insurance for treatments such as chiropractic and acupuncture, based in San Diego.
American Specialty Health's IT budget is allocated by a steering committee consisting of the CFO, the senior vice president of IT, the COO, the senior vice president of Operations and the director of IT. Chevalier's IT steering committee follows a similarly disciplined approach to IT budgeting--adding in a requirement that an IT project generate a positive net present value.
Chevalier gets lots of technology sales pitches but says he has seen "nothing earth-shattering." However, he yearns to make better use of his scarce time. As such, he suggests that the following could help:
* Common technology standards;
* Integration of systems across a company and between a company and its suppliers and customers; and
* Ubiquitous wireless access to data and voice.
Next 'Big Thing' Candidates?
What today's technology may offer is not a revolutionary shift, but tools to make existing tasks easier and faster. Indeed, three exciting IT developments currently on the horizon could help address Chevalier's and other busy executives' need for more time.
Semantic Web. The first is the Semantic Web (SW). No doubt Chevalier would be thrilled if he could type into a computer that he wanted to convene a meeting of his IT steering committee and, within two minutes, receive an email with a date, time and meeting location that worked for all the participants--all without having to speak with a single person.
Such is the promise of the SW--a concept introduced in 2001 by WWW inventor and MIT professor Tim Berners-Lee. The SW is an extension of the current Web. While the Internet exchanges information for people, the SW will mediate information for machines. It's virtually replacing a person with a machine.
To illustrate the difference, compare scheduling a meeting. Today, an executive's administrative assistant sends out an email to all participants suggesting proposed dates. The assistant gets back three replies immediately saying that the proposed meeting times are unworkable. Then, once a satisfactory date is established, other committee members may finally read the original email and agree to the original time; the assistant contacts these other members and informs them of the new time. One of these committee members can't attend and needs to participate via conference call. The assistant books the room, arranges for the conference call and emails the participants. The total time elapsed to plan and book the meeting is, say, two hours.
Using the SW, an individual could type in that he wanted to hold a meeting in the next two weeks, and the SW would book the meeting in less than a minute. With the SW, all the committee members' schedules would be stored on their personal digital assistants (PDAs).
The SW would send out a meeting request to software agents for the PDAs and for the company's available meeting rooms. The SW would negotiate among the committee members' and meeting rooms' schedules until a satisfactory time and location was found. Finally, the SW would book the meeting, arrange audio-visual support and email participants with the details.
The SW could save a "live" person almost two hours, because machine "intelligence" would do the scheduling without a live person's intervention.
Sensor Networks. A second technology, Sensor Networks (SNs), may be closer to happening than the SW, but the productivity benefits may not be as great. Wal-Mart Stores Inc. is trying to create a kind of SN through an initiative announced in mid-2003 to require its top 100 suppliers to install Radio Frequency Identification (RFID) tags on their products by Jan. 1, 2005. RFID tags are small, 25-cent components including a chip, antenna and product information that Wal-Mart could track through in-store and in-warehouse RFID signal-reading machines.
While Wal-Mart didn't meet its January 1 deadline, the benefits--reduced supply chain costs, shelves stocked with what consumers want and less of the rest, as well as diminished theft--make the Wal-Mart initiative an SN test case.
SNs may also help with building automation, industrial monitoring and perimeter security. Berkeley, Calif.-based Dust Networks' SmartMesh technology deploys sensors that collect and transmit physical data for such industries. But developers must address big challenges before SNs become the "next big thing."
RFID and SNs create huge amounts of data, generated by tracking many items' locations over time, explains Jeannie Ross, senior researcher at MIT's Center for Information Systems Research (CISR). Ross believes that SN developers must find the right application and the analytical tools to sift nuggets of insight from torrents of data. A technically feasible application might not make business sense. For instance, a retailer could put an RFID tag on a $1 tube of toothpaste to signal the retailer when the consumer emptied the tube, but this would probably not profit the retailer.
Before SN data can pass across corporate perimeters, settlement systems must be developed, argues Kim Perdikou, Juniper Networks' CIO. For example, if an SN data packet passes through a local telephone company, a cable broadband provider and a private wireless carrier, the participants must agree how to hand off the data and allocate and charge the relevant costs.
Furthermore, SN developers must protect data privacy, notes Bill Dally, Stanford University Computer Science chair. For motorists in traffic, real-time data on the position and speed of vehicles ahead of them on the highway could be more useful than radio traffic reports. But unless an "anonymizer" protects motorists' identities, people worried about their privacy would likely oppose the system.
IBM's On Demand Computing initiative combines the SN and SW to create what Gary Cohen, general manager of IBM's Pervasive Computing Group, calls "a more democratized, inclusive way of thinking about technology that brings together people throughout an enterprise's departments and across its value chain."
Easier-to-Use PCs. A third executive time-saver is to make commonly used devices, such as PCs, much easier to use. In a 2004 Wall Street Journal op-ed piece, Yale Computer Science Prof. David Gelernter argued that the roughly quarter-century-old PC is "a primitive, infuriating nuisance." He even hints about the consequences to the U.S. economic health--of not updating the PC--since IBM has, in essence, given up on further developing the PC with its recent decision to sell the division to a Chinese company. Gelernter believes the PC to be at a juncture where there is plenty wrong with it, but he's optimistic that "there are dozens more possibilities" to make PCs easier to use.
For example, he suggests that email should be streamlined. When a user starts up the computer, a one-page email summary could list emails according to the following categories: Important, Junk and Acknowledged (sent a one-line "I got it" message, but not yet answered). He envisions that a PC could offer an "email summary" built-in, along with a new key that allows a current message to be acknowledged.
Other enhancements he suggests would also make PCs easier to use include:
* Offering "transparent information sharing," which would let users access all electronic documents via the Internet regardless of the machine on which they were created.
* Using inexpensive data storage to eliminate the need to delete email and other electronic documents.
* Using large-screen and projection technology to display menus in a three-dimensional "viewport" that enables users to "peer through" them.
* Providing the exact same layout of the same icons for each computer an individual uses, regardless of the interface.
Stanford's Dally takes Gelernter's notions a step further--from rejiggering software to re-engineering entire work-support systems. He believes that too many systems are designed to satisfy the limits of memory and computing power, and that with the technological advances over the last 20 years, these limitations should cease to drive system design.
Dally suggests that work-support systems should fit naturally with the way people work rather than vice versa, as they have in the past. Instead, he suggests that systems make it easier for people whose work involves, say, drawing diagrams or recording data in notebooks to capture that information on an electronic notepad that would enhance their effectiveness.
Financial executives taking a hard line with their IT budgets are playing a kind of waiting game before loosening their grip on those budgets. However, if technologies like the SW, SNs and easier-to-use PCs described here can save time, then these executives might well decide that such technologies are worth the investments, and will open up their "wallets" and spend more liberally again.
So even if this decade doesn't create an innovative wave, greater technology spending can happen. But, it will likely be more individualized--on a company-by-company basis, and based on what technologies each views as most value-producing.
Total IT Spending by Decade ($ in Billions) 1970s $300 1980s $800 1990s $3,000 2000s $15,000* *estimated Source: Howard Rubin, Meta Group Note: Table made from bar graph.
Peter Cohan is president of Peter S. Cohan & Associates (www.petercohan.com), a management consulting and venture capital firm. His seventh book is Value Leadership: the 7 Principles That Drive Corporate Value in Any Economy (Wiley, 2003).
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||roi of technology|
|Author:||Cohan, Peter S.|
|Date:||Apr 1, 2005|
|Previous Article:||Bring on competition in the accounting profession: the CEO of accounting firm Grant Thornton LLP argues for more competition, saying that depending...|
|Next Article:||Electronic trading: portal to the future? Technological advances are moving markets toward an all-electronic environment, thanks in part to the...|