Using Commitments to Manage Across Units

To coordinate work across different business units, executives should think of the organization as a nexus of commitments, or personal promises between employees, that must be actively managed.

Executives must often manage nonroutine projects, such as integrating company mergers, filling market gaps that fall between current business units, rolling out large-scale IT systems and developing innovative solutions to new customer needs. For such pioneering work, a company’s installed business processes are frequently ineffective because they’ve been designed and optimized to execute routine activities. As such, organizations often experience much difficulty handling novel initiatives, particularly when important work needs to be coordinated across different business units, functional groups or geographic locations.

Problems with horizontal coordination are not necessarily a sign of bad management. Instead, they are nearly inevitable in any complex organization with multiple, specialized divisions. Business subunits adapt and fine-tune their metrics, language and incentives to excel at their required tasks. This very specialization, however, increases the difficulty of integrating activity across units.1 Salespeople talk past manufacturing engineers, and neither understands the scientists in research and development. The standardized work and information flows that cut horizontally across an organization can help coordinate action, but such flows work best when executing routine activities. In fact, their very reliability and efficiency limit their flexibility in doing new things.2 How then should companies perform nonroutine activities that require cross-unit coordination?

From our research, including case work at more than 100 organizations, we have found that managers can better execute novel initiatives across units by viewing the organization as a nexus of commitments — or personal promises that employees make to each other (see “About the Research.”). Using that perspective, they can take practical steps to identify critical commitments within the organization, locate and diagnose breakdowns and intervene to fix them. Managing by commitment increases organizational flexibility because managers and employees can exercise discretion in selecting the best people to work with and negotiate terms tailored to the task at hand. Commitments also increase the effectiveness of execution. Before employees personally commit, they can demand a deeper understanding of what is required and the consequences of failing to deliver. To the extent that they see a promise as a personal pledge, they work harder to honor that commitment.

About the Research »

The Power of Commitments

As we define it, a commitment is a promise made by a performer to satisfy the concerns of a customer within the organization.3 Customer and performer refer simply to roles: An individual acts as a customer when making a request; a performer fulfills that request. A company’s CFO might be a performer when supplying financial data to the IT department but becomes a customer when requesting technical support from that group. In committing to a customer, a performer promises to fulfill the customer’s “conditions of satisfaction,” that is, the specific terms (such as cost, timing and quality) required for the customer to feel satisfied that the performer has fulfilled the pledge. Typically a customer will make a request in order to elicit a promise from a performer, but conversely the performer can take the initiative and proactively propose an offering to meet a customer’s conditions of satisfaction.

To appreciate the power of managing by commitments, consider Imexsa, formerly a state-owned plant that is now the Mexican subsidiary of Mittal Steel Co. N.V.4 When Mittal acquired Imexsa in 1992, the plant was operating at 20% of capacity and losing money. As part of its overall transformation, management initiated a systematic program for making internal service agreements between departments and for monitoring the delivery against those commitments. The head of a department would meet with each internal supplier to articulate his key requirements, explain his rationale for the request, evaluate the agreement against current strategic priorities and agree on concrete conditions of satisfaction. Before promising to achieve the specific targets, a service provider could specify any prerequisites necessary to guarantee delivery.

The head of maintenance, for example, might promise to perform preventative work on machinery within a four-hour window provided being notified one week before the scheduled down time. The performer was responsible for daily monitoring of the task and for monthly customer updates. After an agreement had been executed, the customer would declare whether his conditions had been met. If a performer had repeatedly failed to make good on promises, his or her record would be discussed by the entire team of top managers —something that had happened only once in the system’s first two years.

During that time, Imexsa executed 140 internal service agreements across 28 departments and subunits. Of those commitments, more than four out of five had been fulfilled to at least 95% of the conditions of satisfaction. The result: more effective execution even on novel initiatives, including the adoption of new technology for expanding capacity and the expansion of sales to customers in Asia. Previously, for instance, the manager of maintenance would often receive the wrong equipment from the internal traffic department and, to make matters worse, the delivery would be late, even after he had placed the order two days before the scheduled job and had called three or four times daily to check on its status. Now he gets the right machinery on time, and the traffic department calls him to check on his level of satisfaction. Today Imexsa is profitable, and the plant is one of the world’s largest and lowest-cost producers of high-quality steel slabs.

Commitments That Work

People often contrast speech and action with phrases like “all talk, no action” or “rhetoric versus reality.” But speech act theory argues that talking is doing whenever someone commits to a future course of action by making a promise. Promises are much stronger than descriptive statements, which simply transmit information. For example, the vice president of manufacturing for a pharmaceutical company projects a certain number of batches of a new drug by simply stating a best guess based on current information. But when making a binding promise to deliver a certain quantity, the VP assumes personal responsibility to meet that commitment and to notify those involved if something goes off track.

A reliable promise induces others to depend on it and adjust their behavior accordingly. Such commitments provide the connective tissue that links different activities within organizations. In the pharmaceutical example, the VP of marketing might initiate a series of activities, such as preparing an ad campaign for the new drug, that she might not have undertaken had she been skeptical of her manufacturing colleague’s promise. The notion of reliance is central to the definition of effective commitments. A commitment is effective only to the extent that it generates sufficient confidence in its execution such that customers predicate their own subsequent actions on its successful completion.

A customer will rely on a commitment only when three conditions are met. First, she must believe that she’s achieved a shared understanding with the performer about what needs to be done. (Note: To simplify the discussion, customers will generally be referred to as female and performers as male.) Reaching a “meeting of minds” is inherently difficult in nonroutine activities because customers find it difficult to specify the desired outcome when tasks are complex and require the articulation of tacit knowledge and when new knowledge will emerge as the activity proceeds.5 To complicate matters, customers and performers often speak past each other when their specialized subunits differ in language systems, time horizons or priorities.6 Second, the customer must also believe that the performer has the necessary skills to deliver. Lastly, she must also believe that he is motivated to deliver, specifically, that he will give her commitment a high priority and find the wherewithal to persevere when unforeseen obstacles arise. Motivating performers to execute can be difficult when corporate policies evaluate and reward employees on the basis of their own unit’s performance because such policies undermine incentives to cooperate across groups.7

Satisfying the above three conditions is easier said than done, but our research suggests that the most powerful commitments tend to share the following characteristics: They are public, active, voluntary, explicit and motivated.8 Of course, the effectiveness of any given commitment also will be influenced by other variables, such as the parties’ existing relationship and the performers’ loyalty to the organization as a whole. Nevertheless, our research suggests that these five characteristics define the essence of any commitment.


People consider public promises more binding than private ones, and they strive to avoid any dissonance between their past resolutions and current actions.9 Public commitments also raise the stakes by putting the speaker’s reputation for competence and trustworthiness on the line.10 Moreover, a public promise creates a moral obligation to the extent that others have relied on the commitment to their own potential detriment.11 When a performer doesn’t deliver, he betrays the customer’s trust, particularly if she suffers from his failure.

Public commitments also increase the quality of coordination by broadcasting the intended action to others who can then flag hidden problems or overlooked opportunities. Ideally, a performer makes the promise to the customer in front of his peers as well as his boss, and the group then publicly tracks the progress of the commitment. Many companies, for instance, post the status of various tasks of their cross-functional product teams. Some firms go much further. Consider Haier, headquartered in Quing-dao, China, which has risen from a failing collective enterprise to a leader in China’s home-appliance market in the span of two decades. At Haier, managers’ performance against their commitments is posted publicly, along with rankings of how well they are doing relative to their peers, and this information is updated monthly.


A commitment requires continual communication between the customer and the performer in order to clarify the conditions of satisfaction and to surface any potential obstacles, particularly as unforeseen contingencies arise. Probing discussions are important because people’s background assumptions influence the way they interpret requests and promises, and misunderstandings occur when those backgrounds differ. “Do what’s best for the customer” might mean cheap prices (to a salesperson) or reliable quality (to a manufacturing engineer) or ergonomic functionality (to a product designer). Thus both the performer and customer should take joint responsibility for hammering out the conditions of satisfaction. The performer, for example, should specify exactly what he requires to meet the customer’s needs, including additional resources, relief from other responsibilities and political cover. Such discussions will help avoid mismatches between the performer’s capability and skills and the task’s requirements.

Active discussions also increase a performer’s motivation to honor his commitments. When a customer explains to a performer exactly why a task is important to her, that conversation increases his empathy for her, which in turn deepens his sense of obligation. Furthermore, performers feel a greater sense of ownership in a commitment when they’ve had the opportunity to voice their concerns and requirements.12


A commitment is voluntary only when the performer is free to decline the request. Psychologists have found that people assume greater personal responsibility for promises they’ve made when they are not being pressured. Put differently, when “no” is not an option, “yes” means little. “Voluntary” applies not only to the initial commitment but also to the performer’s ability to renegotiate the conditions of satisfaction at critical checkpoints or when unforeseen events arise. Giving people the authority to decline requests can help them better allocate their limited resources. Often, performers are in the best position to assess the true cost of an activity that may seem insignificant to the customer requesting it. Bringing such costs to light, including the potential risks and obstacles to fulfilling a request, can help customers make better decisions about the best use of an organization’s resources.13

The agile approach to new product development illustrates how voluntary commitments can work in practice.14 In most agile methods, project managers do not assign tasks but instead allow team members to volunteer for activities. Those commitments are then continually tracked and renegotiated, sometimes on a daily basis.


An explicit commitment clarifies who the performer is, what he will do and for whom he will do it. Note that committees do not commit, nor do departments, task forces or teams. That is, effective commitment starts with an individual performer taking personal responsibility for satisfying a customer’s needs. At Mittal Steel, based in Rotterdam, the Netherlands, each internal service agreement always includes the name of the customer and the individual on the line for getting the job done. This helps personalize the commitment: The performer has the name and face of the customer in mind when executing the agreement, and the customer has someone to hold accountable. If the performer fails to deliver, he will suffer a personal blow to his reputation. Explicit promises clear the fog of anonymity that surrounds collective commitments made to abstract goals like “capturing synergies” or “filling the white spaces.”15 When the “who?” “what?” and “to whom?” are explicit, commitments leave little place to hide.

The conditions of satisfaction must be sufficiently explicit so that the performer knows what he must deliver and the customer knows what to expect. In one agile approach to new product development, the customer lists her desired features on small index cards and prioritizes them based on the available resources. Team members then generate a list of tasks required to provide those features, which they then volunteer to accomplish. But even though the conditions of satisfaction must be explicit, they should never be written in stone. That is, the conditions of satisfaction must remain open to renegotiation as part of the ongoing dialogue between customer and performer.


When the customer provides her rationale for her request and the performer understands and accepts it as important, the commitment becomes motivated. The most effective rationales are grounded in a shared understanding of what is important for the organization as a whole. By linking a specific commitment to the company’s overarching strategy and priorities, customer and performer increase the alignment of their individual commitments with the overall corporate direction. By articulating the rationale underlying her request, the customer increases the performer’s understanding of the task at hand. Armed with this deeper comprehension, the performer can then execute the request more effectively. He might, for example, suggest alternative approaches for achieving the same objectives. And when a performer understands why a request is important, he is more likely to elevate it relative to competing priorities.

Commitments Versus Processes

Commitments, of course, are not the only mechanism for coordinating activity across business units. Many managers, for instance, rely on business processes instead.16 The process perspective frames the organization as a bundle of interconnected routines, or standardized tasks and information flows. The commitment perspective looks at the company as a dynamic network of interconnected promises (see “A Comparison of Process- and Commitment-Based Management”).

A Comparison of Process- and Commitment-Based Management

View Exhibit

Process-based management helps improve the efficiency, quality and speed of routine activities such as order fulfillment, manufacturing and customer service. Using various techniques, including Six Sigma and Total Quality Management, managers are able to dramatically reduce costs and resource inefficiencies. But process-based management also constrains an organization’s flexibility to respond to unforeseen contingencies and to execute nonroutine activities.

Commitment-based management, on the other hand, enhances a firm’s ability to handle nonroutine tasks. Internal customers and performers enjoy great latitude in defining the conditions of satisfaction, which allows them to generate creative solutions to new problems. They can also modify those conditions in light of new information or shifting priorities. And customers can cast their nets widely to identify the best performers for particular projects, which further enhances the organization’s flexibility. The personal nature of a commitment, moreover, acts as the engine that drives delivery. Also, renegotiating a promise is generally less cumbersome and time-consuming than the reengineering that is often required by process-based management.17

But managing by commitments has limitations as well. The greatest risk is that customers and performers will agree to side deals that are locally optimal but poorly aligned with the company’s overall strategic priorities. Furthermore, identifying the best performer and negotiating the conditions of satisfaction take considerable time (although that up-front investment can easily be recouped if it helps avoid problems down the line). Customers must understand performers’ capabilities and constraints; performers must learn customers’ needs; and both parties must negotiate workable conditions of satisfaction, monitor progress and renegotiate as necessary. Finally, because commitment-based management is a new approach, it currently lacks the breadth of supporting infrastructure that process management offers, including frameworks, tools, certification bodies and consultants.

Consequently, we propose commitment-based management as a complement to — not a substitute for — process-based management. Many companies can benefit from the use of process-based management to continuously improve routine activities. But when an organization faces totally unprecedented initiatives, commitment-based management can provide numerous advantages, and it can enhance the effectiveness of many existing processes. For instance, the approach helped the senior executives at a U.S. biotechnology firm move faster on licensing deals. Previously, the company had missed valuable opportunities because of a lack of cooperation among various departments, but commitment-based management helped improve the teamwork among those groups. Now the head of preclinical research shares information about promising antibodies early in the process with the director of business development, who is then much better able to explore the commercial development of those specialized proteins with potential partners.

Generating Effective Commitments

Commitments emerge out of ongoing discussions between performers and customers. Those interactions generally proceed through four basic steps — preparation, negotiation, execution and acknowledgment — with each step often requiring multiple conversations between customer and performer.

Most managers have an intuitive grasp of the importance of the four steps, but few of them understand what explicitly needs to be done in each. As a result, they rush through them, sometimes even skipping steps, or they devote all their attention to one step, virtually ignoring the others. But a breakdown in any step can easily trigger a downward spiral. After a customer picks the wrong performer for a job, for example, she grows frustrated and begins to intervene. But her micromanaging only annoys the performer who slowly begins to disengage from the project. Eventually, he checks out altogether. The customer then burns herself out trying to do the performer’s job in addition to her own.

To avoid such pitfalls, customers and performers must understand what they need to do throughout all four steps. And, perhaps more importantly, they should be aware of the hidden dangers in each.


The customer first prepares a request, articulating what she wants accomplished and why it matters in the larger strategic context. She then identifies the right performer for the job. Although the importance of these tasks might seem painfully obvious, many customers rush through them, and the result is that they end up selecting the wrong person, typically with disastrous consequences.

A customer unwittingly lays the groundwork for later trouble when she fails to clarify what she wants and why. In such cases, customers fall into the trap of the proverbial military officer who requested a rock from a subordinate without describing the rock or offering a rationale. When the soldier returned, the officer replied dismissively, “not that kind of rock, another rock.” To avoid that trap, the customer should pause to resolve in her own mind what she wants — and why — before making a request. Clarifying that rationale against corporate priorities provides a simple check to ensure strategic alignment.

A common mistake in selecting a performer is to choose the person who happens to occupy the designated position on the organization chart, ignoring that individual’s skills or current capacity. Instead, customers should interview a potential performer just as they would a job candidate, or they should conduct a quick background check with colleagues who have worked with the performer before to inquire about that person’s skills, current priorities and workload, as well as his potential enthusiasm for the project in question. The preparation step ends when the customer makes a request to a performer (or a performer makes an offer).


Here, both parties clarify the conditions of satisfaction, and the performer specifies what he will need, not only in terms of resources but also with respect to relief from current responsibilities and even political cover. People often shortchange this stage (or omit it altogether) because they assume that the other party is thinking along the same lines, but that can be a prelude to disaster. Consider the design engineers at a global construction firm that spent more than £400,000 to produce a bid for a power plant. The problem was that their customer (in this case, the company’s sales team) wanted the plant to showcase the firm’s environmentally friendly technology, while the lead engineer had assumed that such “green issues” were secondary to technical elegance.

Negotiation can be time-consuming and complex, but it is an essential step for clarifying the performer’s current priorities, constraints and potential obstacles. The give-and-take between customer and performer helps define the “hard edges” of the conditions of satisfaction. Customers who hurry through this stage often end up securing passive promises that lack the performer’s in-depth understanding and personal buy-in.

To specify the conditions of satisfaction, the performer must understand the rationale for the request and know exactly what constitutes success, in particular, the terms that matter most to the customer, such as cost, timing or quality. Once that much is clear, the two parties can discuss critical contingencies that could alter the customer’s request or stymie execution. To check whether both parties are in sync, the customer could ask the performer to restate the conditions of satisfaction and the underlying rationale for the request.

Managers often assume that people will say “no” to any request that doesn’t come directly from their bosses. But that’s hardly the case and, surprisingly, a meaningless “yes” is a far more common problem than an automatic “no.” In fact, people often feel compelled to say “yes” to maintain their reputations as team players, to avoid conflict and to stay in the good graces of senior executives. Such ceremonial commitments have little power to obligate the performer to the customer.

Actually, getting to “no” requires conscious effort. Performers need the political space to decline customers’ requests or propose a counteroffer that is consistent with their skill level, available resources, time constraints, backlog of jobs and political obstacles. Instead of an insincere “yes” or a flat-out “no,” the goal is for the response from a performer to be along the lines of, “What you’re asking is not possible, but this is what I can do for you” or “Yes, provided you can do X, Y and Z for me.”

Although organizational politics are often a taboo subject, customers need to discuss such issues explicitly in order to take responsibility for clearing any potential roadblocks. At a leading U.K. retail bank, for example, a marketing team had developed an interactive computer program that allowed branch managers to analyze a person’s financial situation. The system was designed to attract new clients, but the branch managers were under enormous pressure from the CEO to sell more loans. So the managers instructed the sales force to use the new application to persuade existing clients to consolidate all of their loans with the bank. That hard sell, however, only alienated people. In this case, the head of marketing had the responsibility of convincing the CEO to relax the pressure he was exerting on the branch managers and allow the program to be used as the marketing team intended.


Both performers and customers often mistakenly view this step as the point at which they can stop talking and get to work. Nothing could be further from the truth. In fact, ceasing to communicate (or switching to “radio silence”) is as dangerous in the execution step as it is in the negotiation step. Customers and performers must continue to monitor progress, discuss upcoming obstacles and emerging opportunities, and renegotiate the conditions of satisfaction as circumstances shift. Through such an ongoing conversation, both parties can adapt to various unforeseen contingencies, such as changes in prices, competitors’ tactics and corporate priorities.

Radio silence is the source of many breakdowns in execution. The customer often equates silence with progress and assumes the performer is on track. Or the customer equates silence with inaction, so she begins to micromanage. On his part, the performer interprets silence as a sign that the customer’s needs and priorities are stable — generally a bad assumption. By failing to communicate until the scheduled delivery date, the customer and performer practically ensure that unpleasant surprises will crop up at the last minute. To avoid that, many companies use a formalized system to monitor critical commitments. For example, America Latina Logistica, a logistics company based in Curitiba, Brazil, publicly tracks its managers against critical commitments by using red, yellow and green markers to denote each performer’s status.

Renegotiation during execution is rarely pleasant, but delaying or avoiding it usually only leads to further difficulties. Performers should renegotiate as soon as they realize they cannot achieve any critical condition of satisfaction. Prompt disclosure enables the customer to adjust and mitigate the damage. The customer also has an obligation to initiate renegotiations if her priorities or circumstances shift in ways that affect the conditions of satisfaction. Otherwise, she could place the performer in a no-win situation. At a large bank, for instance, the CFO asked the CIO to consolidate the services offered to line managers into “factories.” When the business-line leaders complained, the CFO withdrew support for the initiative but he failed to inform the CIO, who soldiered on until the organizational resistance became overwhelming. At that point, the CIO quit in frustration, setting the entire project back a year.

Any renegotiation should be conducted with both parties acting in good faith. If, for example, the customer initiates a renegotiation, the performer should not seize on it as an opportunity to wriggle out of his original commitment.


In the final step, the customer declares the commitment fulfilled (or not). Like an umpire calling a pitch either a ball or a strike, the customer has the unilateral power to determine whether her conditions of satisfaction were met. In this regard, the customer is always right. Although this step should be the easiest, it, too, is prone to breakdowns.

A common problem occurs when the performer delivers without inquiring about the customer’s satisfaction, and the customer grudgingly accepts the work without giving any feedback. Failing to complete the acknowledgment step might seem like a trifling matter, but it deprives both customers and performers of critical information that could help them better manage their future commitments. Also, when customers fail to declare their satisfaction publicly, they deny performers recognition for a job well done.

People shy away from public acknowledgment because they often fear giving or receiving feedback. Customers avoid giving feedback, even when it’s constructive, because they’re afraid of alienating valued colleagues. And, of course, preparing and giving constructive feedback requires time and effort. On their part, performers dislike inquiring about customer satisfaction because they fear that any criticism will tarnish their reputations. Surprisingly, many organizations lack the institutionalized processes through which their internal customers can publicly declare their satisfaction when projects are completed. Annual 360-degree feedback sessions, for example, seldom coincide with the fulfillment of particular commitments.

When an organization does have established feedback practices, they often exist primarily as mechanisms for customers to vent their dissatisfactions. One major U.S. credit-card issuer, for example, conducted feedback conversations as audits, triggered only when performers made mistakes. Not surprisingly, employees developed a strong aversion to the process. To be effective, public declarations of satisfaction (or dissatisfaction) and the associated feedback should coincide with completion of the commitment, rather than waiting for an annual review. They should also be conducted face-to-face, and they need to show the effect of the fulfillment on the business. Lastly, they should always include praise (whenever appropriate) as well as critical feedback.

COMMITMENTS PROVIDE a simple yet powerful lens to help understand how company initiatives get done (or why they stall). The framework is effective for coordinating work across business subunits, but its applicability extends to a broad range of challenges near the top of many CEOs’ agendas, including making out-sourcing arrangements work, increasing the effectiveness of existing processes and providing integrated customer solutions. But that’s not to say that all companies should uproot their processes and replace them with commitment-based management. Commitments do, however, provide a fresh perspective on management as well as a set of practical tools and techniques for improving an organization’s flexibility, execution and performance when handling some of the most intractable challenges that executives currently face.


1. See P. Lawrence and J. Lorsch, “Differentiation and Integration in Complex Organizations,” Administrative Science Quarterly 12 (1967): 1–30.

2. See R.R. Nelson and S.J. Winter, “An Evolutionary Theory of Economic Change” (Cambridge, Massachusetts: The Belknap Press of Harvard University, 1982); and M.T. Hannan and J.H. Freeman, “Structural Inertia and Organizational Change,” American Sociological Review 49, no. 2 (1984): 149–164.

3. We use the words “promise” and “commitment” interchangeably. Our core argument extends to a range of promises including those made to individuals outside a firm’s boundaries and budget commitments made inside a firm, among others; but we focus on horizontal commitments across organizational subunits in this article. A commitment resembles an implicit contract, which plays a central role in several streams of theory, including agency theory: see M.C. Jensen and W. Meckling, “Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure,” Journal of Financial Economics 3 (1976): 305–360; transaction-cost economics: see O.E. Williamson, “The Economic Institutions of Capitalism” (New York: Free Press, 1985); and psychological-contract theory: see D. Rousseau, “Psychological Contracts in Organizations: Understanding Written and Unwritten Agreements” (Thousand Oaks, California: Sage, 1995). We share with these literature streams a focus on agreements between parties — whether they are called “implicit contracts,” “psychological contracts” or “commitments” — as an important alternative to processes for coordinating activity within organizations and across their boundaries. These research streams have, by and large, focused on a specific dyadic relationship as the focus of analysis, for example, between an external supplier and customer in transaction-cost economics, between principal and agent in agency theory and between senior executives (as representatives of the firm) and employees in psychological-contract theory. We differ from this literature first in using commitment as a general construct that can be used to analyze a broad range of promises. More importantly, our focus is the ongoing discussion required to make effective commitments rather than the specific content of the agreement.

4. D.N. Sull, M. Escobari, T. Sugata and J. Cabera, “Spinning Steel Into Gold: The Case of Ispat International N.V.,” European Management Journal 17, no. 4 (1999): 379.

5. Activities are nonroutine to the extent that individuals cannot fully specify in advance the desired outcome or most effective process to achieve that outcome. Difficulty in specification stems from three sources: complexity arising out of a large number of variables and interactions among them, tacit knowledge required to articulate the desired outcome or optimal approach, and emergent knowledge that arises during the process. Nonroutine activities also include initiatives that can be specified but are not customary in a specific context.

6. Recent scholarship has distinguished between problems of cooperation (agents know the optimal action but lack incentives to pursue that action) and problems of coordination (agents have incentives to cooperate but don’t know what to do). See C. Camerer and M. Knez, “Coordination, Organizational Boundaries and Fads in Business Practices,” Industrial and Corporate Change 5, no. 1 (1996): 89–112; and C. Camerer and M. Knez, “Coordination in Organizations: A Game Theoretic Perspective,” in “Organizational Decision Making,” ed. Z. Shapira (Cambridge, U.K.: Cambridge University Press, 1997). Joint action across specialized subunits often poses both cooperation and coordination challenges.

7. Collaboration across organizational subunits can be viewed as a public good or a resource that all parties benefit from regardless of whether they have contributed to its creation or maintenance. Public goods create a temptation for employees to obtain a free ride and benefit from the fruits of collaboration without helping to create or maintain them. In multiunit organizations, the free-rider problem is aggravated when multiple parties are involved. Employees can free-ride with greater anonymity and spread the cost of their free-riding across many people as the number of individuals involved increases. See P. Kollock, “Social Dilemmas: The Anatomy of Cooperation,” Annual Review of Sociology 24 (1998): 183–214.

8. Gerald Salancik argued that commitments in general are binding to the extent that they are explicit, public, voluntary and irrevocable. See G.R. Salancik, “Commitment and the Control of Organizational Behavior and Belief,” in “New Directions in Organizational Behavior,” eds. B.M. Staw and G.R. Salancik (Chicago: St. Clair, 1977): 1–54. We agree with Salancik (and a large body of empirical evidence from social psychology) that individuals consider their promises more binding to the extent that they are public, explicit and voluntary. Revocability applies to a set of actions — Salancik uses the examples of shooting a friend or having a vasectomy — that are not generally relevant for speech acts, which are rarely irrevocable. We add the characteristics active and motivated because they apply to promises (a subset of all commitments) for reasons we elaborate on in the text.

9. See L. Festinger, “A Theory of Cognitive Dissonance” (Stanford, California: Stanford University Press, 1957) and C.A. Kiesler, “The Psychology of Commitment: Experiments Linking Behavior to Belief” (New York: Academic Press, 1971).

10. B.R. Schlenker, D.W. Dlugolecki and K. Doherty, “The Impact of Self-Presentations on Self-Appraisals and Behavior: The Power of Public Commitment,” Personality and Social Psychology Bulletin 20 (1994): 20–33.

11. For discussion of the ethical obligations arising from promises, see J.R. Searle, “How to Derive ‘Ought’ From ‘Is,’” The Philosophical Review 73 (1964): 43–58.

12. Psychologists have found that individuals feel more obligated by their commitments to the extent that they actively discuss them prior to making a promise. The investment in effort required to clarify the conditions may also contribute to the strength of perceived obligation. For a recent review, see R.B. Cialdini, “Influence: Science and Practice” (Boston: Allyn & Bacon, 2000), 52–97.

13. Research on intrafirm strategy-making processes has found that distributing resource allocation decisions throughout levels within large, complex organizations allows them to better harness specific knowledge dispersed throughout the firm. See J.L. Bower, “Managing the Resource Allocation Process” (Boston: Division of Research, Harvard Business School, 1970) and R.A. Burgelman, “A Process Model of Internal Corporate Venturing in the Diversified Major Firm,” Administrative Science Quarterly 28 (1983): 223–244.

14. For a comprehensive overview of agile project-management methods, see J. Highsmith, “Agile Project Management: Creating Innovative Products” (Reading, Massachusetts: Addison-Wesley, 2004) and C. Larman, “Agile and Iterative Development: A Manager’s Guide” (Reading, Massachusetts: Addison-Wesley, 2003). Commitment-based management shares many features with agile project management, particularly the emphasis on satisfying the customer and the importance of frequent, face-to-face conversations.

15. One can think of commitments as a mechanism for converting the interunit coordination problem from a multiplayer game to a two-person game, which simplifies both the incentive and information challenges required for effective collaboration.

16. T.H. Davenport, “Process Innovation: Reengineering Work through Information Technology” (Boston: Harvard Business School Press, 1993) and M. Hammer and J. Champy, “Reengineering the Corporation: A Manifesto for Business Revolution” (New York: HarperCollins, 1993). For a recent review of process research, see D.A. Garvin, “The Processes of Organization and Management,” Sloan Management Review 39 (summer 1998): 33–50.

17. G. Hall, J. Rosenthal and J. Wade, “How to Make Reengineering Really Work,” Harvard Business Review 71 (1993): 119–131. Since 1993 Darrell Rigby of Bain & Co., a global management consulting firm, has conducted a survey of the use of management tools. The 2003 survey revealed that reengineering ranked 19th out of 25 tools in terms of usage by firms, 20th in terms of satisfaction and fourth in terms of customers abandoning the tool. See D. Rigby, “Management Tools Survey 2003,” Strategy & Leadership 31 (2003): 4–11.