CMP Media / Optimize - Business Strategy & Execution for CIOsCMP MediaOptimize
Optimize Magazine DisciplinesExpertsGap AnalysisDiscussionsWhite PapersExecutive Programs Information Week Optimize Magazine
Optimize Magazine
Optimize - Ideas. Action. Results.

ROI Valuation
The IT Productivity GAP
horizontal dots
By Erik Brynjolfsson
July 2003, Issue 21
Why the ongoing debate over whether IT contributes to productivity growth? While productivity—the amount of output per unit input—isn't the only thing a business or an economy has to worry about, in the long run, it encompasses just about everything. Ultimately, productivity growth is what determines our living standards, the competitive advantage of companies, and the wealth of nations. It's arguably the single most-important economic statistic.

For Americans, the good news is that while the numbers on the budget deficit, the trade deficit, unemployment, and overall economic growth have been disappointing for the past couple of years, productivity growth has remained remarkably strong. In fact, last year, according to the U.S. Bureau of Labor Statistics, the annual rate of output per worker grew at 4.8%. This was almost double the rate of the previous five years, which in turn, was nearly double the 1.3% growth rate of productivity in the 1980s and early 1990s.

My ongoing research, based on data from more than 1,167 large U.S. companies, finds a statistically significant correlation between the intensity of IT used in a company—IT capital per worker—and that company's overall productivity. There's an emerging consensus among economists that IT has been the biggest single factor driving the productivity resurgence, although debate continues about the exact magnitude of its contribution.

From the chart, it's easy to see that the overall relationship between IT and productivity is positive. However, it's also evident that there's tremendous variation in performance among companies. Each point represents a specific company in the data set. IT is measured as the current replacement cost of IT hardware stock per worker. Productivity is measured by the concept of "total factor productivity," which is defined as real output divided by a weighted average of all inputs, including labor and non-IT capital. Each measure is normalized by industry (two-digit SIC code), with the average for the industry defined as zero. (For more details and the research paper describing our findings.)

The Kmart/Wal-Mart difference

The critical question facing IT managers today is not "Does IT pay off?" but rather, "How can we best use computers?" Even when their IT intensity is identical, some companies have only a fraction of the productivity of their competitors. Clearly, IT isn't like a certificate of deposit where you can invest your money and expect a guaranteed rate of return. Managers at Kmart won't catch up with Wal-Mart simply by installing new IT systems. While effective IT use is an important part of Wal-Mart's success, IT has been the catalyst for a broader host of changes in information flow from Wal-Mart's customers all the way to its suppliers. Only by understanding these IT-enabled changes and matching or even improving them can other retailers hope to close the productivity gap.

The same is true in every industry: IT is only the tip of a much larger iceberg of complementary investments that are the real drivers of productivity growth. In fact, our research found that for every dollar of IT hardware capital that a company owns, there are up to $9 of IT-related intangible assets, such as human capital—the capitalized value of training—and organizational capital—the capitalized value of investments in new business-process and other organizational practices. Not only do companies spend far more on these investments than on computers themselves, but investors also attach a larger value to them (see chart, p. 37).

Working with Lorin Hitt of the Wharton School, Tim Bresnahan at Stanford University, Shinkyu Yang at New York University, and colleagues and students at MIT, I've been systematically studying the relationship among IT, productivity, and organizational practices.

Our research began with visits to dozens of companies, many of them sponsors of the MIT Center for eBusiness, to learn more about how they generate value from their IT investments. For instance, at Cisco Systems, we observed an emerging "Internet culture" of distributed information flow, worker empowerment, and ubiquitous access to Web-based data for employees, suppliers, and customers. At United Parcel Service of America, we saw the importance of end-to-end integration of systems and a focus on execution. At Dell Computer, we witnessed the transformation of the factory floor using new production-planning systems and a dramatic reduction in work-in-process inventories. Merrill Lynch showed us how difficult it can be to transition from traditional work practices to those that fully embrace and leverage the Internet, but also how much shareholder value such a transition can generate. British Telecom worked with us to refine a software tool called the Matrix of Change, which facilitated the transformation at BT and some of its largest customers.

Next, we studied whether the practices were idiosyncratic or part of a broader pattern. We then conducted a series of interviews with more than 300 companies asking about work practices. The new data was matched with previous data on IT investments and performance metrics, including productivity and market value.

The results were telling. We found that the greatest IT benefits are realized when an IT investment is coupled with a specific set of complementary business investments. Companies that use IT intensively work differently from their competitors.

Most businesses already know that computers aren't an end in themselves. While the thousandfold improvement in the price/performance of information technology over the past 30 years is a credit to the industry, IT creates value only if it lets users work more effectively. This is especially true for information workers, who now constitute the majority of the U.S. workforce.

Think of a company as an information-processing organism. It gathers data from the market, suppliers, and employees. It makes decisions, implicitly and explicitly. It then communicates and acts on those decisions, with some degree of proficiency. For instance, Wal-Mart uses data from its point-of-sale system to decide which products to stock and passes this information on to suppliers so they can improve their production planning. Similarly, Dell acts on customer information to build and ship custom products within 24 hours of receiving an order. However, many other companies become paralyzed when their internal systems handle too much data.

Too often, the flow of information speeds up dramatically in highly automated parts of the value chain only to hit logjams elsewhere, particularly where humans must get involved and processes aren't updated. The end result is little or no change in overall performance. A gigabit Ethernet network does no good if the real bottleneck is a manager's ability to read and act on the information. In the information economy, the scarce resource is not information, but the capacity of humans to process that information.

To increase a company's information metabolism, it's not enough to simply automate parts of business processes or even automate whole processes without considering how the rest of the organization will be affected. Business processes and decision-making systems are inevitably linked in myriad ways. Most business processes, such as the expense-approval system or hiring process, evolved during a time when information-processing costs were radically higher. Given the unprecedented reduction in costs over the past 20 years, it's not surprising that the decision-making structures that were optimal in the 1970s and 1980s aren't still optimal today.

Our analysis of Wal-Mart and Kmart showed that Wal-Mart had consistently higher penetration of network technology, as measured by network nodes per worker. But it also had significantly different decision-making systems in place—key decisions, such as the authority to approve certain purchases, were more decentralized and distributed at Wal-Mart. This combination of technology and organization was correlated with higher levels of productivity and market value. Comparing data from 1997 and 2002, we found that Kmart shifted its technology and decision-making systems in the direction of Wal-Mart, but Wal-Mart also shifted, maintaining its differentiation on all three measures: technology, organization, and performance.

Overall, we found that IT is consistently correlated to changes in the way people work and how their performance is measured, controlled, and reported. For instance, centralized databases provide individual workers with the necessary information to complete an entire process that historically was fragmented. This shifts a worker from the role of functional specialist to process generalist. In manufacturing, the use of flexible machinery and computerized process controls is often coupled with greater worker discretion, which in turn requires data-analysis skills and general problem-solving ability.

In general, simple decisions closely related to individual transactions or other operational actions, such as back-office applications, have been most amenable to computerization. More complex and cognitively demanding work—such as that which dominates the time of managers and professionals—still proves remarkably difficult to automate. Computer automation of even clerical and blue-collar work typically doesn't directly substitute for all of a worker's tasks, but instead replaces a subset of ancillary tasks.

We also know that the role of computers is different from the way traditional machinery ushered in the industrial age when machines replaced muscles. While it's sometimes popular to think of technology as "automating information work," this isn't a good analogy. Rather than replacing human brains, computers complement most cognitive tasks and actually increase the demand for human information processing. Computer technology, at least at the current state of the art, is a relatively poor substitute for the mental capabilities of people. In fact, we found that companies that are intensive and effective IT users tend to employ more skilled, educated, and highly trained workers than their peers.

Highly computerized processes, such as online customer service, are often accompanied by a greater production of data. In turn, raw data is fodder for analytic or abstract decision making—such as analyzing customer needs to target new-product development. The net result is a heightened value for skilled workers, managers, and professionals. These workers are also higher paid. The premium paid to superstars with exceptional skills has been growing most rapidly of all, and that's a reflection of the organizational shift that we observed in the most productive users of IT.

Even as companies attempt to hire more skilled and educated workers, they must also make organizational adaptations to distribute information-processing tasks. For example, even with advanced data-analysis tools, no top manager can take full advantage of all the information available to a company. In fact, the problem of information overload and the resulting bottleneck of human information-processing capacity tends to worsen at the higher levels of the corporate hierarchy.

As a result, the most productive IT users were more likely to find ways to empower line workers and salespeople, and otherwise decentralize decisions. Technologies such as intranets, wireless networks, and information-sharing software not only distribute decision-relevant information throughout the organization, but they also provide an infrastructure for an effective incentive system.

Interestingly, we find that the changing nature of information work is also placing greater demands on noncognitive skills, such as management. Employees at all levels need more skills in dealing with customers and suppliers, influencing teammates and colleagues, and inspiring and coaching subordinates—in other words, the "people skills" that computers lack.

It's clear that new technologies and organizational structures require more flexibility and autonomy than in traditional employee roles where the production process is fixed and less discretion is needed. The productive companies in our sample were adjusting to these changing requirements.

All these changes and variables point to the fact that companies don't simply plug in computers or telecommunications equipment and achieve productivity gains. Instead, they go through a process of organizational redesign. Our statistical analysis found that IT is embedded in a "cluster" of related innovations, notably organizational changes outside the IT department. This cluster, which we call the digital organization, includes:

  • Automation of numerous routine tasks.
  • Highly skilled labor.
  • More decentralized decision making.
  • Improved information flow vertically and laterally.
  • Strong performance-based incentives.
  • Increased emphasis on training and recruitment. On average, companies that successfully combine these elements produce more valuable output than competitors and achieve high levels of productivity. They also tend to have higher employee and customer satisfaction.

    IT hardware can be purchased, but implementing the digital organization requires a more difficult process of "co-invention" by IT users. For example, Merrill Lynch struggled for more than two years to develop a strategy that took advantage of the new ways the Internet let it interact with customers.

    Identifying and implementing organizational co-inventions is difficult, costly, and uncertain, yielding both successes and failures. Our data showed that these adjustment difficulties—including financial constraints, obsolete work rules, and even incompatible corporate culture—led to significant variations in the use of IT and the resulting outcomes. Ironically, the difficulty that so many companies have in implementing the practices of the digital organization is exactly what creates the opportunity for extraordinary returns among the successful implementers. If implementing these practices were easy and straightforward, all competitors in an industry would already have done so, driving their competitive advantage to zero. Today, adopting an ATM network is no longer correlated with a competitive advantage for a bank, but adopting the practices of the digital organization is.

    The main conclusion is that in advanced economies, IT is a promising source of productivity growth, but it makes little direct contribution to the overall performance of a company or the economy until it's combined with complementary investments in work practices, human capital, and organizational restructuring.

    We do find evidence of a substantial relationship between computers and productivity growth, but closer examination reveals that the biggest benefits accrue to companies that adopt an identifiable cluster of business practices we call the digital organization. Relatively few of these practices are directly related to the implementation of the technology itself. Most involve changes in the organization of information work, including decision-making rights, incentive systems, hiring, and training.

    Not all companies, even among aggressive users of IT, have successfully implemented these practices. But companies that are unusually productive have overcome the adjustment costs associated with organizational innovation. They've adopted a new cluster of information work practices.

    Erik Brynjolfsson is the Schussel Professor of Management at the MIT Sloan School of Management and director of the Center for eBusiness at MIT.

    See a comparison of human strengths vs. machine.

    The 90-Day Plan

    Investment in IT capital alone doesn't guarantee productivity. Our research reveals organizational characteristics that can be implemented to form a coherent system of complementary practices within a company.

    FIRST MONTH: Develop measures of IT, productivity, and organizational characteristics

  • "You can't manage what you don't measure" the saying goes, and that's especially true of IT and productivity. Before deciding in which direction to head, know where you are. Begin by assessing the IT capital stock already in place, including equipment and software purchased by departments and users, as well as equipment outsourced but used by the company. Be sure the IT department can provide a complete and accurate picture of the IT capital the company owns and uses.
  • Document the outputs and inputs of each division, including labor and capital costs. Companies with multiple plants, locations, or groups that provide similar outputs—for example, retail branches, factories, and regional offices—have a particularly good opportunity for benchmarking against one another.
  • Identify and document the investments in work practices that are important for your company, including the decision-making systems, incentive systems, training investments, and other business processes. This is the most difficult step because it varies by industry and strategy goals and requires data not readily available.

    SECOND MONTH: Compare practices with effective IT users

    Just as IT investment creates a capital good, so too does "investment" in organizational practices and culture. Statistical analysis shows that certain characteristics are correlated with productive use of IT, including:

  • Embedding standard procedures in technology so employees and customers can work with less supervision.
  • Distributing decision rights and empowering line workers through decentralization and delegation.
  • Offering strong performance-linked incentives, such as individual performance-based incentive pay.
  • Emphasizing "human capital," including hiring highly educated people and providing ongoing training. There are important exceptions in the data—not all successful companies adopted these practices and not all adopters were successful. In fact, adopting the practices indiscriminately can hurt performance. Consider whether they make sense in your situation. Prioritize appropriate practices.

    THIRD MONTH: Adopt a change-management process

  • Changing organizational practices and culture can be difficult and risky. Change-management tools, such as the Matrix of Change, can be helpful in navigating this change.
  • Enlist help from executives, managers, and employees to implement and maintain the practices you identify.
  • Use the company intranet to communicate productivity goals and objectives.

    Sidebar Online Only: Human vs Machine

    Copyright © 2003 CMP Media LLC
  • Optimize Magazine
    Optimize Magazine Optimize Magazine Optimize Magazine Optimize Magazine Optimize Magazine
    Optimize Magazine Optimize Magazine
    Optimize Magazine