Governance

Outlook Special Edition, October 1999

Governance is a major source of alliance underperformance. Our 1998 cross-industry survey shows that a number of common causes of alliance failure are related to poor governance (See Chart 1). In the oil industry, for example, our interviews indicate that 78 percent of executives believe governance to be a major cause of alliance failure. And even more tangibly, poor alliance governance costs the pharmaceutical industry somewhere between $5 billion and $18 billion per year.

What is alliance governance? Why is it the source of so much difficulty—and what can companies do about it?

We define governance as the process of oversight that both provides for efficient, informed interchange between two or more allied companies and brings accountability to those who run an alliance. More specifically, governance has four primary goals: to protect the interests of the corporate parents, to allocate resources needed to make the alliance succeed, to arbitrate conflicts among lower levels of staff and to help extend a venture's upside potential beyond initial projections.

Most industries have developed a single, standard model of governance. In the oil industry, for instance, it has been to treat alliance governance like corporate governance: A board of directors provides distant oversight of venture management, protecting the financial interests of nonoperating shareholders. In other industries–such as pharmaceutical and automotive, where nonequity alliances are common—companies treat alliance governance like project management, creating a steering committee that effectively serves as a discussion forum for all parties involved.

In both instances—quasi-corporate and quasi-project—companies have tended to use a single governance standard for all their alliances. This is reasonable enough, so long as the alliances being governed by that model are alike.

A Systematic Approach
These days, however, more and more corporate portfolios are filled with alliances that differ dramatically in terms of goals, duration, resource contributions and value. Royal Dutch/Shell, for example, is currently involved in more than 10 fundamentally different forms of alliances. Obviously, one model for alliance governance is not going to work for so many different forms of collaboration.

Perhaps even more important than variety is the trend toward invasiveness. In the past, most alliances were either peripheral to the core business of the corporate parents or were central to the one parent serving as venture operator. But many of today's alliances gain value from accessing the core skills of each corporate parent.

These are invasive alliances, and their ranks include the Swissair-Austrian Airlines alliance and the British Telecom-AT&T joint venture. These alliances demand that the parent companies manage all sorts of exposure and risk, as well as a much freer flow of resources, especially tacit know-how and staff. At the moment, there is no precedent to rely on for governing such alliances.

Governance by a single body, board or committee is not an option for most alliances, especially invasive or high-value alliances. The flow of activity is simply too complex or important to rely on a single group. A better model is a governance system, in which the governing process is conducted through various levels of decision-making bodies that may include CEO summits, boards of directors, steering committees, operating committees, alliance managers and various project committees.

The example of General Electric and Honeywell shows how this works in practice. Initially, the companies created a board of directors to handle the governance of their industrial controls joint venture. Over time, realizing that the interaction was too extensive to be overseen by the board, the partners established an operating committee and various project committees. Many networks, including the international airline alliance Star Alliance and SEMATECH, a US-based semiconductor research consortium, have recognized the inherent complexity and have established governance systems from the outset.

The best governance systems will include four practices:

1. Vary Governance Style, Tailoring it to the Alliance. Under a system, the shape of governance can vary according to certain attributes of the alliance. For example, a short-term co-marketing alliance will need a governance model unlike that of a $1 billion new business joint venture. Many forces influence the shape of an alliance-governance system, including venture goals, number of partners and past relationships between the companies. Two elements, however, are particularly crucial: alliance value and contributed resource complexity.

2. Clearly Assign Rights and Responsibilities. Governance systems introduce new problems. Most common among these is a poor definition of roles to be played by the various levels of governance. When a governance system is hazy in dividing responsibilities, it will suffer from internal conflict and slow decision making. To avoid this problem, the corporate parents should formally define not only which decisions each governance level will be involved in but also the nature of that participation. This should occur in the late stages of negotiation.

It is a three-step process. First, a task force appointed by the executive sponsors within each corporate parent should list the 15 to 20 most important decisions that the alliance will need to make, including the annual plan, staffing, transfer prices and dividend policies. Then this group will need to create a decision-making spectrum—a method that demarcates the various ways in which different individuals or groups can be involved in a decision. One biotechnology alliance, for instance, divided participation rights into five classes: propose, review, veto, decide and monitor. Finally, the task force has to assemble these elements, identifying

which governance levels will be involved in which decisions and the manner in which they will participate. This process allows each governance level to understand the range (and limits) of its authority over the alliance.

3. Monitor and Support Information Flows. Another common problem is poor information flows. In alliances today, most information is nonnumerical and must pass through various hands. This combination of traits makes information hard to transfer, easy to misinterpret and, frankly, prone to disappearnce

As a result, companies will need to create a detailed accounting of what information each governance level will need, determine the source of that information and make the oversight of its flow an explicit task for those involved. Some large alliances may benefit from information support teams. Chevron and Texaco have each established a small corporate team (known as affiliate analysis units) to support the board members of Caltex, their $17 billion-per-year downstream joint venture. These groups ensure that the joint venture board members get the information and analysis they need to make decisions.

4. Create Clear, Telling Performance Measures. Although a governance system is a unique organism, it is not so unique that its performance cannot be measured. Measurement is critical for ongoing success and self-correction. One useful gauge is "pass-ups"—decisions that one governance level had the power to make, but passed upward to a more senior group. High pass-up volumes are a symptom of a dysfunctional governance system. They indicate that decision-making rights were inappropriately allotted or, more commonly, that an unnatural, potentially destructive nervousness pervades the governance system. It follows that pass-up volumes can be used to measure the effectiveness of governance. A national co-locating and co-marketing alliance between a major oil company and a fast-food restaurant chain made the number of pass-ups a prime gauge of alliance performance, tracking the number of pass-ups and then linking staff incentives to the month-to-month decline in this number. The companies found that monthly pass-ups dropped from more than 100 in the first month to fewer than 10 after six months. The result was that the alliance was faster to market than a similar alliance without such a measure. Given the importance of governance in alliances, companies should create other measures of governance effectiveness in addition to pass-ups. Possible measures include: Speed on major decisions (benchmark against internal business unit performance). Member turnover rates, especially at the board level. Director participation levels (percent of members attending meetings). Extent of governance dovetailing (degree to which one member of each level is an insider to the level above it). Performance ratings of individual board members. Degree of alliance expansion (changes in scope indicate active governance). Executives should look at each governance system and determine a set of measures that are consistent with fulfilling the vision of the alliance. Alliance governance is no longer a simple or uniform task. In the future, companies will need innovative governance systems to oversee their alliances, working to ensure that each system reflects and supports the goals of the particular alliance. .

Back to Contents


To Top