Why new technologies fail.

Technology implementations are expensive and often fail. Failures - cases in which new technologies perform below expected levels - are estimated to run between 40 percent to 75 percent. The Standish Group's 1996 study of 360 companies indicated that 42 percent of corporate information technology projects were abandoned before completion. Their 1994 study reported that three-quarters of enterprise resource planning system installations were judged unsuccessful by the companies paying the bills. Another Standish Group study of 365 information technology executives found that 48 percent felt there were more failures in 1994 than the five years prior. A 1998 survey of 376 CEOs by CSC Index and the American Management Association reports that 50 percent of all technology projects do not meet CEOs' expectations.

Clearly, technology implementation success could be improved with active top management support, clear implementation goals, and user involvement and training. Unfortunately, continuing high failure rates suggest little widespread application of these implementation techniques.

Too often, managers charged with implementing a new technology receive an edict similar to Star Trek Captain Jean Luc Picard's command, "Make it so, Number One," with little in the way of time, money, or active support for the implementation effort. We believe that perceptual biases and existing organizational systems mask critical reasons for supporting the implementation process, which makes implementation the invisible partner in technology introductions. In this article we first explain why the advice offered by technology implementation models, though well documented, is often overlooked or discounted. Second, we offer several concrete suggestions on how implementation project managers and teams can overcome these biases and barriers to stimulate support for the implementation process. Our analysis is developed from a variety of sources, including surveys, structured interviews, published examples, and experimental data.

The invisibility problem

Implementation includes any process undertaken to institutionalize a new technology as a stable part of an organization. Implementation follows the adoption decision and is bounded by institutionalization, in which the technology becomes part of the status quo. Unfortunately, implementation is the clear underdog in a new technology introduction. For instance, a major information systems text describes the design phase as challenging and creative, while parts of the implementation phase are described as "unsexy afterthoughts ... second-class jobs that 'fit in' with the rest of the process." Similarly, articles advocating cost models to justify technology acquisitions rarely mention the cost of managing technology implementation. This relative invisibility is reflected in the beliefs and actions of both executives and implementation managers in many organizations.

The invisibility of implementation results in a tendency for senior managers, who make the decision to acquire major new technologies, to overestimate the value of a new technology and the likelihood of implementation success. Superficially, the estimated value of a successfully implemented technology may seem easy to establish. Executives usually have substantial information from the vendor about the full-capacity value of the technology they are buying. The full-capacity operation of a robotics assembly system, for example, may be rated at four times the output per labor hour compared to the system it is replacing. Hence, the value of the robotics system can be calculated. However, it is rarely the case that complex new technology implementations immediately achieve full capacity.

Caterpillar, for example, took more than two years to get its first larger, flexible manufacturing system to function effectively and several more years to achieve full realization of the system's capabilities. General Motors labored over its flexible manufacturing system for more than 18 months after installation without achieving its estimated capabilities. And five years after their construction, continuous slab casters in U.S. steel mills continued to operate at only half their rated capacity.

Executives also seem to project that their efforts (even minimal ones) are likely to result in the stated benefits of the technological change. The historical reality of 40 percent failures has not been enough to adjust managers' views. Just as most of us believe we are above-average drivers, most executives seem to believe that they, and the people who work with them, are above-average change agents.

M. Lynne Markus and Robert Benjamin make an even stronger statement based on their experiences with information technology specialists and consultants involved in information technology-enabled organizational transformation projects. Markus and Benjamin note that managers often have a magic-bullet theory when it comes to information technology-enabled transformation: They have built the gun and the magic-bullet of information technology cannot do anything but hit its mark. Since the bullet will assuredly hit its mark, there is little reason to worry about who is going to aim and fire the gun. The researchers note that "The magic-bullet theory is seductive to many IT specialists and line managers because it allows them to disembody change ideas, package them as technologies, and distance themselves from the hands-on sport of helping people to change."

These are costly perceptions to hold. Thirty years of theory and research on implementation and change management show that active top management support is a major factor in successful implementation. Executive attention and support are critical for successful implementation because resources for change generally flow from the top down. Current models of implementation and change management provide many ideas about the direction that should be taken and types of support that need to be provided when implementing new technology. However, as Richard M. Nixon noted, "It's not enough for a leader to know the right thing. He [or she] must be able to do the right thing."

Somewhere below senior

managers in the hierarchy, project managers recognize the importance of implementation issues but often do not act on that knowledge. Consider the case of 10 technology projects being implemented in a Fortune 300 manufacturing organization. Project engineers were asked to describe the budgets of their projects. These projects had total budgets ranging from $20,000 to $2 million for a variety of technologies (e.g., bar code systems, plastics extrusion lines, and high-speed product identification marking). When asked how important "people issues" were to the success of their projects, project engineers ranked them from critical to moderately important.

Compare their responses to the information presented in Figure 1, which shows the percentage of the budgets that were spent on implementation items, such as training, informational literature, or events to promote the technologies. Nine of the 10 projects used two percent or less of their budgets on people issues (and out of these, five reported that they budgeted no funds for people issues), while one engineer estimated that his project had allocated 10 percent to such items. Clearly, there is some disconnect between the importance of issues such as providing adequate training and the money allocated to assure success.

PMs understand that the implementation process is important, yet their budgets do not suggest that they follow through on their understanding. One problem lies in how budgeting takes place. In this organization, for example, when asked about training, information, and literature that was budgeted for a technology with a price tag of $150,000, the project manager noted: "All these come out of 'operations' budget ... 'non-productive time.'" In other words, this project manager was counting on the everyday operating budget to cover the time and costs of training and information for a new technology. The budgeting process often only looks at tools and equipment; it does not include opportunities to budget for people-related issues. In this organization, in fact, training and other people items would be counted as a negative mark against the project engineers' performance. Any resources used in that area would be perceived as a large cost.

The bias against budgeting for people issues is not constrained to this firm. Traditional financial accounting practices treat equipment as assets and people as expenses. Buying a technology is typically an external event with an externally set price. Implementation, on the other hand, is internal. Accounting systems make it much easier to acquire funding for something with an externally determined value that can be depreciated. Time off for training, payment of internal consultants, commitment-building activities, and the like are much more difficult to fund. These internal costs have no hard value, which decreases the visibility of implementation.

However, even when financial analysis takes people issues into account, there still appears to be a bias against investing in implementation. We were able to demonstrate this effect by providing hypothetical implementation data to two groups of masters-degree students. When no return on investment information was available for money spent on hardware versus implementation (training and support) or when the ROI for hardware and implementation were equal, most students would invest the bulk of their budget in hardware. Even when the ROI was 70 percent return for implementation and only 50 percent for technology, only a small majority of the students would invest more into implementation than hardware. Surprisingly, this bias was less pronounced for the students in the MIS masters program than for those in the MBA program. We asked the respondents if they had an explanation. The MIS respondents suggested that the technology is always changing and thus the hardware investment is fleeting. The MBA respondents suggested that the people are always turning over and thus the training investment is fleeting.

Making implementation visible

PMs need to take an active role in reducing the invisibility of implementation, and they can do so in three ways.

1. Reframe expectations. The perceptual biases discussed earlier create problems for PMs by making the implementation process invisible, and what is hard for executives to see is hard for them to support. One key to unlocking top management support is reframing executive views regarding the likelihood of implementation success without an investment of their time and resources. PMs can increase the visibility of implementation by using data about the ease, costs, and outcomes of implementation of a specific technology in other organizations and with information about their own organization's experience with change more generally.

PMs should start with the technology's vendor, particularly if the PM has not been part of the decision to adopt the new technology Information from the vendor can create a clearer sense of top management's expectations for the new technology However, PMs need to keep in mind that the information presented by vendors may be biased. Vendors may emphasize the benefits of a technology and downplay its costs in order to ensure users' initial interest and to rationalize their presentation of the technology. A lack of balanced information about costs and benefits can reinforce the magic bullet view of technology implementation.

Balance can be gained by collecting information from other organizations that have attempted to implement this or a similar technology. Find out from the technology's vendor who else has gone through the implementation process and talk with their PMs. Ask questions such as:

* What kind of utilization rates were you able to achieve after three months, six months, and one year?

* How long did it take your organization to get the technology running at full capacity?

* What types of savings were achieved in labor, materials, and other costs in three months, six months, and one year?

* How much did your organization spend on implementation-related activities such as job redesign, restructuring, and team-building activities? How much do you think your organization should have spent?

* What kinds of training were done and what training costs were incurred by your organization in preparing the labor force to operate the new technology?

* What other implementation costs were incurred?

* What questions should you have asked before the implementation process was underway but didn't?

This type of information can help PMs create realistic expectations of the gains that can be achieved and increase the likelihood that implementation costs are not invisible.

The second set of data PMs need to collect is about their own organization's implementation track record. PMs may discover that their organization is so good at implementation that a top management edict of "Make it so" is warranted. More likely, however, they will discover that there have been both failures and successes in the past. Highlighting what was learned from past implementation efforts can prove useful in arguing for implementation funding and can help produce a realistic time frame for implementation. Find out what implementation efforts the organization has undertaken during the past few years. Then ask:

* What happened to those implementations?

* What support and funding did they receive from management?

* To what extent did those implementation efforts encounter resistance in the organization?

* What problems were encountered?

* What did other PMs within your organization learn from their experiences?

The ideas stimulated by this process can provide a realistic notion of how different techniques affect implementation success. PMs can use the information gained to develop estimates of the value of various levels of success as well as the real costs of failure, providing them with more persuasive measures of the value. By making the connection between implementation efforts and the realization of a technology's value, PMs should be in a stronger position to argue for implementation funding and support.

2. Create small wins. Evidence of early implementation success can be an effective technique for influencing executives. Early success is more likely if small, easy steps are attempted first - and data collected and disseminated about the outcomes. We suggest that PMs plan for a series of small wins rather than trying to pull off one major coup. Moreover, smaller projects are more likely to succeed overall. The Standish Group's chairman says that when project costs (one measure of project size) reach $10 million, the chances of the project coming in on time and on budget are "statistically zero." Large-scale change efforts often fail because they are conceived on too big a scale.

A comparison of two automobile manufacturing plants provides an illustration of these concepts. A plant using teams focused on improving quality globally with no time frame or other goals had, at the end of six months, generated hundreds of suggestions and abundant good will - but little in the way of improved quality A plant with a sharply focused two-month plan of reducing the most prevalent defect within one process reached its goal and then went on to address other sharply focused changes successfully This strategy of small wins can be applied in a technology implementation setting. By achieving a series of smaller goals, rewards and reinforcements come frequently, and visible progress makes the linkage between the implementation effort and technology success explicit.

Similar results have been found in banking. Banco Itamarati claims prominence as one of the first banks to use client/server architecture to its full advantage. Its objective was maintaining a close relationship with customers by offering financial solutions and customer satisfaction, all to the benefit of the Itamarati Group. The company produced a 51 percent annual net profit growth and moved from 47th to 15th place in the Brazilian bank ratings. Managers give three fundamental reasons for their success: (1) clear vision with documented, specific objectives; (2) top-down involvement; and (3) the ability to produce incremental, measurable results throughout the planning and implementation period.

In contrast, Washington state's motor vehicles department was not so fortunate. Its client/server drivers' license processing system was scrapped after $40 million was spent in the course of five years. The system was initially budgeted at $16 million and scheduled for a three-year start-up. Todd Saner, assistant state director of information systems, notes that the project was conceived on too large a scale and that they have now learned to keep projects small and incremental.

The moral for PMs is clear. Do not define successful implementation as complete utilization of the new technology Create the understanding that implementation is an ongoing, adaptive process that moves forward step by step. Demonstrate success by taking on achievable projects with explicit objectives that focus on generating results instead of completing activities. For example, providing training in support of a new technology implementation is not a concrete implementation goal. Rather, what can users do at the end of the training session? What kind of outcomes can they produce?

3. Reduce conflicts of interest. Many organizations try implementing new technologies in the existing organizational setting, which can create problems for PMs. Existing organizational structures and processes tend to treat new technologies as invisible because they were not designed with the technology in mind. Numerous studies show that technology benefits are more likely to be gained when the introduction of technology is accompanied with other organizational process improvements. For example:

* The installation of SAP (the leading enterprise resource planning software suite) derives 80 percent of its benefits from changing the business process; the software is just an enabler.

* Research on advanced manufacturing technology implementations shows that up to 60 percent of the anticipated gains in some organizations were achieved through structural changes made prior to the actual implementation of the technology.

Existing structures and processes can create conflicts of interest. While implementation may be in the best interest of the organization, it may not be in individual users' personal interests. One of the clearest examples is when a new technology is implemented on top of an existing organizational evaluation and reward system. Middle managers, for example, may see the benefit of a new technology for the organization but still resist its implementation if it is not in their own interest because of existing performance criteria.

End-users often face the same dilemma and resist implementation because the technology conflicts with their personal interests. Dorothy Leonard-Barton's research found that even though structured system analysis provides a stronger background for software engineering, use of such systems is sometimes seen as having a negative effect on individual performance. In one case, the programmer and analysts felt they were judged chiefly on whether or not they completed projects on time. Given that the analysis required up-front time in the development process, requests to use such techniques were often ignored.

Such problems can be dealt with by modifying current performance criteria. These modifications can range from large-scale changes in an organization's evaluation and reward system to local variances in existing criteria to specifically support the implementation effort. Consistent with a "small wins" strategy, initial changes in evaluation and reward systems should be handled on a variance or experimental basis because they are less likely to trigger resistance that way Moreover, keeping changes small and localized reduces the need for active top management support early in the implementation process. The success of early small changes will build top management support for the implementation process and make more substantial change possible later.

Other areas PMs should assess include existing job roles, reporting relationships, and decision-making authority These are as are usually fine-tuned to support the existing technology, and the roles needed to support new technology implementation often do not fit within the existing structure. For example, during the implementation of a laser measuring system into an auto assembly line set one inspector took on the role of supporting the new technology (by keeping it calibrated and learning uses for the system). Union grievances were filed, as these roles did not clearly fit into the inspector's job category. Additionally, the other inspectors did not support the time and training necessary for the one inspector's new role, creating tension within the group as well as grievances across groups. PMs should look for ways to make small, localized changes in roles, responsibilities, and decision-making authority As in the case of performance criteria, small changes are likely to encounter less resistance, reducing the need for early top management support, which, in turn, increases the chances that early successes will generate support for later changes.

Visibility enables implementation

The relative invisibility of the implementation process in new technology introductions is a major cause of implementation failure, and such failures are costly to both the organization and the people involved. Creating an understanding of the linkages among implementation funding, support, and technology success can be one of the most difficult perceptions for PMs to construct. Executives must understand that real resources - money, time, and attention - are paramount to the successful use of the technology. Without these resources, all but the most necessary changes will fall below expectations. With the support and resources needed to train and socialize personnel, reward achievements, create commitment, and adjust structures and procedures, however, implementation success is much more likely Paradoxically, it seems that executives must see success before they will support implementation.

So when the edict to "Make it so" comes from above, PMs need to think through the issue of how to create a context that enables implementation by generating support. The key to motivating support for implementation and effectively using the detailed advice provided by the technology implementation management literature is found in increasing implementation's visibility.

Creating a context that supports implementation is PM's first responsibility in the implementation process. Without the support that visibility generates, there is little opportunity to use the tools and techniques provided by the implementation literature. And without those tools and techniques, successful implementation is a long shot.

For further reading

Benjamin R.I. and Levinson, E., "A Framework for Managing IT-Enabled Change," Sloan Management Review, Summer 1993.

Griffith, T.L. "Negotiating Successful Technology Implementation: A Motivation Perspective," Journal of Engineering and Technology Management, 1996.

Markus, M.L. and Keil, M., "If We Build It, They Will Come: Designing Information Systems That People Want to Use," Sloan Management Review, 1996.

Zammuto, R.F. and O'Connor, E.J., "Gaining Advanced Manufacturing Technology's Benefits: The Roles of Organization Design and Culture," Academy of Management Review, 1992.

The Authors

Terri L. Griffith, Ph.D., is an associate professor of organizational behavior and technology management at the John M. Olin School of Business, Washington University. Her research and consulting interests include the implementation and effective use of new technologies. She is a member of the Academy of Management and INFORMS, serves on the editorial boards of the Journal of Engineering and Technology Management, Group Decision and Negotiation, and the Journal of Managerial Issues, and is a senior editor for Organization Science.

Raymond F. Zammuto, Ph.D., is a professor of management at the University of Colorado at Denver. His current research focuses on the roles of organizational culture and new technologies in facilitating or impeding organizational change. He is a member of the Academy of Management and INFORMS and is an editorial board member of Organization Science.

Lynda Aiman-Smith, Ph.D., is an assistant professor of business management in the College of Management at North Carolina State University. Her research and consulting interests are in implementation issues, especially implementing new technology or implementing teams into organizations. She also researches organizational recruiting of technical personnel. She is a member of the Academy of Management and the Society for Industrial Organizational Psychology.