This is an MIT Sloan Management Review article. Uses two different studies--a survey of CIOs at 256 enterprises in the Americas, Europe, and the Asia/Pacific region and a set of 40 interview-based case studies at large companies such as Johnson & Johnson, Carlson Companies, UPS, Delta Air Lines, and ING DIRECT--to conclude that when senior managers take the time to design, implement, and communicate IT governance processes, companies get more value from IT. Toward that end, they offer a single-page framework for designing effective IT: a matrix that juxtaposes the five decision areas (principles, architecture, infrastructure, business application needs, and prioritization and investment decisions) against six archetypal approaches (business monarchy, IT monarchy, federal, duopoly, feudal, and anarchy). Illustrates how successful companies use different approaches for different decisions to maximize efficiency and value for both IT and the overall enterprise. Offers recommendations to guide effective IT governance design.

Strategic alliances continue to be important tools of competitive strategy. However, many alliances do not achieve their partners' collaborative objectives and are terminated prematurely. As a result, alliances are often described as inherently unstable organizational forms that are subject to high rates of failure. In addition to the problems associated with not achieving collaborative objectives, a related issue has received limited attention. Although alliances are prone to failure, there are numerous examples of strategic alliances that continue for years despite failing to accomplish partner objectives. This article examines such strategic alliances. However, it does not focus on why failure occurs, but on the variables that contribute to firms' persistence with failing alliances. It also provides measures that can effectively counter persistence.

Managers often take projects well past the point at which they should drop them. To see if they have come to this point, managers must look closely at themselves and recognize which of the influences they may be under. Some influences are psychological--they've been rewarded in the past for sticking to their guns, so why shouldn't the same thing happen this time? Some are social--no one likes a loser. And some are structural--important members of the organization are publicly committed to the project. The rest of the job belongs to top management. Its course is to rethink what behavior it rewards and how it staffs projects and to ensure that its information systems report the real odds.