Why IT Does Matter

Harvard Business Review editor-at-large, Nicholas G. Carr, ignited a firestorm in the opinion piece “Why IT Doesn’t Matter” published in the May 2003 issue of HBR.

Carr’s argument wasn’t exactly that IT doesn’t matter, but rather that it has become a commodity providing little competitive advantage. As a result, he said, companies should rethink how much they pay for IT given this reduced return on investment.

HBR received a large number of positive and critical responses to Carr’s piece including a letter we offer here from two professors at Harvard Business School. —Ed.

In no other area is it more important to have a sense of what you don’t know than it is in IT management. The most dangerous advice to CEOs has come from people who either had no idea of what they did not know, or from those who pretended to know what they didn’t. Couple not knowing that you don’t know with fuzzy logic, and you have the makings of Nicholas Carr’s article.

Carr’s examples of railroads and electric power played out over eighty years, (not forty, as he suggests), turning society, business organizations, and lifestyles inside out. The deeper societal impacts came during the second forty years, as society’s insights on how to use the technology changed. It is worth noting that although these technologies mutated significantly (for trains, it meant moving from fifteen miles an hour to eighty miles an hour), the mutation was on a totally different and much smaller scale than IT’s.

The cost performance of IT technologies over the first forty years changed by roughly 107, and for the foreseeable future will continue to evolve at the same rate. That is in sharp contrast to a train, which after eighty years moved six times faster than it had in the earlier period. This is impressive, but not nearly as dramatic as a computer produced in 2000, which runs 10 million times faster than a 1960s’ computer.

Carr’s graph on information technology stands as a subject lesson for Darrell Huff’s well-known book How to Lie with Statistics. Carr’s chart would look very different if he had tracked the number of MIPS or CPU cycles on the network from 1990 to 2002. Even using a log scale on the vertical axis would be barely enough to tilt a vertical straight line enough to create something resembling the curves of the other two schematics in Carr’s article. With this explosion of cost effectiveness has come the ability to do things truly differently. American Hospital Supply’s distribution software and American Airlines’ SABRE reservation system are examples of victories in past technologies. The firms were the first in their industries to see technology’s transforming potential, they had the courage to invest in its performance, and they used it to gain a significant competitive edge. It is naive to assume that other sharply discontinuous technologies will not offer similar transformation opportunities in the future.

In our view, the most important thing that the CEO and senior management should understand about IT is its associated economics. Driven by Moore’s Law, those evolving economics have enabled every industry’s transaction costs to decrease continually, resulting in new economics for the firm and creating the feasibility of products and services not possible in the past. The economics of financial transactions have continually dropped from dollars to cents. New entrants have joined many industries and have focused on taking strategic advantage of IT’s associated economics. Company boundaries have become permeable, organic, and global in scope through IT networks and the Internet.


As the pace of doing business increases, the CEO and senior management team must be aware of how IT can change rules and assumptions about competition. The economics of conducting business will likewise continue to improve—providing opportunities for businesses to expand the customer value proposition by providing more intangible information-based services. For example, the automobile value proposition continues to expand with technology that continuously senses road conditions and applies the appropriate wheel traction and suspension system pressures.

CEO and senior management must understand that historical constraints of every kind continue to be knocked off IT because it is a “universal information-processing machine.” Before e-mail and the Internet, the cost of communications was seen as limiting IT’s wider use. Packet switching was invented as a way to digitize voice, data, and video in a matter that enabled digital computers (and its associated economics) to communicate, and the cost of communication sharply and suddenly dropped. Similar situations have transpired with the advent of digitized photography, use of radio frequencies for various handheld IT appliances, and the development of such products as elevators that call in to the service center or to a computer that automatically dispatches collective software or people when a part or system is about to fail. Often, only the senior management team’s imagination limits new IT-based opportunities.

Our research suggests the following:

New technologies will continue to give companies the chance to differentiate themselves by service, product feature, and cost structure for some time to come. The first mover takes a risk and gains a temporary advantage (longer if there are follow-on possibilities). The fast follower is up against less risk but also has to recover lost ground. Charles Schwab versus Merrill Lynch and Walgreens versus CVS are examples of this playing out over the past decade. Our advice to the CEO is to look at IT use through several different lenses. One lens should be focused on improving cost savings and efficiencies. Another should be focused on the incremental improvement of organizational structure, products, and services. Still another should be focused on the creation of strategic advantage through extending competitive scope, partnerships (customers and other parties), the changing of the rules of competition, and the provision of new IT-based services to extend the customer value proposition.

Unless nurtured and evolved, IT-enabled competitive applications, like many competitive advantages, don’t endure. Even historic strategic systems like American Hospital Supply’s (after a decade of financial malnourishment) may wind up turning into a strategic liability. Others, however, like American Airlines’ SABRE have shown extraordinary robustness and have permitted the survival of otherwise doomed organizations.

Evaluating these opportunities as well as thinking through their implications and timing, is vitally important, nonboring work. The new technologies will allow new things to be transformed in nonlinear ways. Radio-frequency identification devices for grocery stores, smart cards, and automated ordering systems for hospital physicians are all examples of new process targets that technologies will soon address. In the more distant future we will see the improved creation of drugs and treatments through the ability to rapidly and more deeply analyze huge databases. Understanding the potential and then deciding when the time is right to seize these transformative applications will be neither routine nor boring for the CEO or CIO.

Grid computing, standardization of components, and open systems, far from stifling differentiation, provide a stable platform to build on and offer new ways of differentiating, either by cost, structure, product, or service. Just as literacy stimulated innovation, so do open systems and grids. Outsourcing the commodity infrastructure is a great way to control costs, build competence, and free up resources, which can be used to combine data bits in creative ways to add value. Relatively bulletproof operational reliability will be a key part of the price of success. Back-office or server farms, help desks, and network operations will be outsourced to specialists to attain this reliability (at rock-bottom costs). Packages like SAP further help remove commodity maintenance activities and allow firms to better analyze customer information and provide service at the sharp end. The package of skills needed inside an organization is changing very fast for competition in the information age.

The jobs of the CTO and CIO are and will be of unparalleled importance in the decades ahead. Max Hopper of American Airlines and Paul Strassmann of Kraft and NASA are not the last of a dying breed of dinosaurs, but prototypes of the leadership skills needed for survival.

If you take 1955 (with the IBM 701) as the start date and use eighty years as a technology cycle, 2035 may not be far off the mark for playing much of this out. Even then, the special recombinant nature of this technology makes us uncomfortable calling an end date. We wish Carr were right, because everyone’s golf handicap could then improve. Unfortunately, the evidence is all to the contrary.

How IT Shapes Top-Down and Bottom-Up Decision Making

What determines whether decisions happen on the bottom, middle, or top rung of the corporate ladder? New research from professor Raffaella Sadun from Harvard Business School finds that the answer often lies in the technology that a company deploys. Key concepts include:

  • Enterprise Resource Planning software is a decentralizing technology: It provides information that enables lower-level managers to make more decisions without consulting their superiors.
  • By the same token, Computer-Assisted Design and Computer-Assisted Manufacturing software creates a situation in which the plant worker needs less access to superiors in order to make a decision.
  • The better the data network, the easier it is for workers to lean on superiors and rely on them to make decisions. It’s also easier for executives to micromanage and keep all the decisions in the corporate office.
  • Trust is also a key factor in determining whether decisions are centralized at headquarters or decentralized at the local level. Research finds that the average level of trust of a multinational’s home country tends to influence the level of decentralization in that company.

What determines whether decisions happen on the bottom, middle, or top rung of the corporate ladder? New research offers a surprising conclusion: The answer often lies in the technology that a company uses.

Information-based systems, such as Enterprise Resource Planning (ERP) software, will push decision-making toward the bottom of the corporate ladder. Communication systems, such as e-mail and instant messaging applications, will push the decision-making process toward the top.

And that means developing an IT strategy isn’t all about deploying the best technology, says Raffaella Sadun, an assistant professor of strategy at Harvard Business School.


“The bottom line is that whoever is in charge of the acquisitions and the IT strategy, they obviously cannot just think about the technology side, they also have to think about the organizational side,” she says. “Traditionally, technology is thought of as a tool that enables empowerment, but that’s not always the case.”

Sadun discusses the issue in “The Distinct Effects of Information Technology and Communication Technology on Firm Organization,” a paper she cowrote with Nicholas Bloom of Stanford University and Luis Garicano and John Van Reenen of the Centre for Economic Performance, London School of Economics.

“Technologies that make the acquisition of information easier at the lower level of the hierarchy are associated with a decentralization of the decision-making process,” Sadun says. “On the other hand, we have the communication technologies, which actually do exactly the opposite.”


Companies, however, often fail to consider the disparate roles of their software systems, let alone their effects on organizational behavior. Rather, they lump “information technology” into one amorphous idea—the “IT” department—which encompasses all the technology in the organization.

“Technology tends to be dumped into a single category,” Sadun says. “The reality is that IT is a huge, heterogeneous set of technologies.”

Similarly, when examining issues such as organization and productivity, industry and academic studies historically tend to treat information and communication technologies as “an aggregate homogeneous capital stock,” according to the paper. To that end, Sadun and her fellow researchers set out to show how—and why—managers need to consider the very different organizational effects of communication and information technologies.

“This difference matters not just for firms’ organization and productivity, but also in the labor market, as information access and communication technology changes can be expected to affect the wage distribution in opposite directions,” their paper states.

The researchers looked at non-production decisions such as capital investment, new hires, and new product plans. Such decisions are either centralized near the top of the corporate ladder or decentralized and delegated to the top of a particular business unit. And the decision makers often depend on ERP software, which facilitates the dissemination of information throughout a large company, enabling detailed coordination among various operating units.

Next, they looked at production decisions, which involve figuring out the tasks necessary to meet the goals and deciding how to pace them. These decisions are generally the bailiwick of either a factory floor worker or a supervisor. For those cases, the researchers studied the role of Computer-Aided Design (CAD) and Computer-Aided Manufacturing (CAM) software in decision-making.

In both instances, the researchers hypothesized that the information software would lead to decentralized decision-making. Because the software eases access to the information necessary to make important choices, both the ERP and CAD systems would increase the likelihood that plant managers and production workers would make decisions and act on them without having to consult an executive at headquarters.

On the other hand, the team hypothesized that a rise in leased lines and corporate intranets would lead to a rise in centralized decision-making at the top of the corporate ladder.


In the past, communication often depended on faxes, overnight delivery services, “snail mail,” or site visits. Even with phone calls, it was difficult for anyone at headquarters to make educated decisions and communicate them to branch offices. In those cases, it was natural to cede control of daily operations to a local manager.

With today’s networking technologies, it’s easier for top executives to keep a constant flow of communication with branch offices. However, the network may actually deter innovation. When technology makes it easier to communicate, erstwhile independent workers may find themselves pestering their bosses with e-mailed questions throughout the day. Micromanaging executives find themselves making all the decisions and constantly sending mandates down the corporate ladder.

“Whenever there is a reduction in the cost of transmitting information, it’s easier for the person down in the hierarchy to communicate with the CEO,” Sadun says. “And the CEO can monitor constantly what this person is doing and just give orders, rather than rely on the judgment of those below.”

The research team evaluated data from some 1,000 manufacturing firms in eight countries, including detailed technology rollout histories and surveys that gauged the relative decisional autonomy of plant managers and floor workers. (In gauging the factors that determine whether a firm adopts any given technology, the researchers considered geographic variables that might affect the cost of acquiring the technology—the firm’s distance from the Walldorf, Germany, headquarters of ERP market leader SAP, for instance, and the fact that telecom industry regulations vary from country to country, which means networking prices vary, too.)

The findings were consistently parallel with the hypotheses: An increase in the penetration of ERP systems led to a substantial increase in plant manager autonomy. A CAD/CAM deployment raised the likelihood of floor worker autonomy. But communication technologies served to lower autonomy, meaning more decisions happened at the corporate level.

“I was reassured and surprised at the same time that these results were holding across countries and industries,” Sadun says.


That said, Sadun notes that technology is hardly the only factor that determines whether a firm allows decision-making both up and down the corporate ladder. Another major factor lies in cultural differences across and within countries. In a separate study, Sadun found that otherwise similar companies showed huge differences in decision-making tactics, according to their geographical location. In the paper “The Organization of Firms across Countries,” coauthored with Bloom and Van Reenen, she documents that firms located in areas with high levels of trust tend to be systematically more decentralized than those in areas with low levels of trust.

Sweden and Portugal, for example, seem to be on opposite ends of the trust spectrum. “There’s huge cross country heterogeneity in the way even apparently similar firms decide how to allocate decision rights within the firm,” Sadun says. “Take Swedish manufacturing companies, for example. You see that they are completely decentralized, and the middle manager is basically a mini-CEO with loads of decision-making power. And then you take a firm that produces exactly the same good, but instead of in Sweden, it’s in Portugal. And there, the middle manager doesn’t decide anything and is completely dependent on the authority of the CEO.

“In our research,” she continues, “we argue that different levels of trust are a key determinant of these differences. If a CEO can trust his senior managers, he will be more willing to decentralize decision-making. For example, there might be a lower concern about the fact that managers will use their power to pursue their personal interests instead of those of the firm.”