New Rules for Teamwork

Not that long ago, teams were typically composed of people with similar skills working in the same place. Their efforts were based on the idea that by working together in a well-managed process, they could deliver replicable results. They primarily operated with efficiency and effectiveness but would frequently encounter obstacles, such as internal siloes or bureaucratic processes, that hindered their overall performance. In response, they began to innovate how they worked together, establishing the groundwork for a more flexible, cross-functional approach to their rituals and practices. From the Skunk Works engineers of the 1950s to Brazilian football in the 1960s to the aviation and surgical collaborations that delivered stunning improvements in the 1990s and early 2000s, organizations have continually searched for ways to enhance their teams’ effectiveness.

Today, companies of all types are called on to demonstrate integrated, cross-functional, project-based teamwork in their operations. In the tech world, for example, software engineers are expected to connect with sales teams, data scientists to work on personnel projects, and user-experience designers to partner with plant managers. Team members may be working together not only from different offices but on different continents. Some industries seem to have cracked the code on high-functioning teamwork—software and manufacturing organizations in particular. Other industries, such as those that are more knowledge-based, are playing catch-up.

Build Better Teams:

Series reprint

download PDF

buy copies

Traditional thinking around how to innovate teamwork has often been based on intuition and observation, with a dash of psychology. But increasingly we face an environment of constant change and disruption, and new ideas about how to create well-functioning teams are emerging; some are based on experience, some are guided by new practices, and some are being made up on the fly. None of this has yet cohered into a systematic approach to improving how teams work.

In recent years at McKinsey, we’ve devoted ourselves to developing a new science of teamwork for our organization. This approach allows us to understand how our 4,000-plus teams are performing and to intervene when necessary. It combines the best of established wisdom with new data-driven techniques and insights. It relies on testing, learning, analysis, adaptation, and improvement—in real time and with accountability—to enable continuous learning. And it includes metrics that link practices to outcomes, for both individuals and teams. We have used the framework to test hypotheses about how teams work best and to apply insights to improve performance. We have applied this work not only internally but also at several of the companies that we advise, building on previous learning to understand what effective, cross-functional teamwork looks like in different contexts.

In this article, we present three key principles of our approach that can help teams in any organization perform at their best. And we describe—through examples and a set of action plans how to put them into practice.

[  1  ]

Develop an Operating System

We use the term “operating system” to mean the building blocks for the way team members collaborate, create change, and support one another. Effective operating systems vary widely, depending on the needs and norms of the organization. What they all have in common is that they set out a view of how teams create value, what teams are supposed to achieve, the technical skills each team member is expected to contribute, the processes by which the work will be managed, and the cultural norms and mindsets of constructive collaboration that will guide behavior. The best operating systems embed an ethos of continuous improvement throughout the organization, not just in a single team or department. They are structured enough to provide consistent guidance but loose enough to accommodate changing conditions, priorities, data, and needs.

The hands-on world of auto manufacturing produced one of the earliest and best-known examples of a cross-functional team operating system: the Toyota Production System. Developed in the 1950s, it combined standard work routines with performance monitoring and metrics that flagged issues and opportunities. Frontline managers were trained to collaborate closely with teams from a range of disciplines to resolve problems and drive continuous improvement.

In the 1970s and 1980s, as the software industry started to take shape, the idea of a team operating system moved from the physical to the digital world. Developers began using signals to track the quality and performance of their code, enabling them to adapt quickly when their work generated an error. Ultimately, in the late 1990s and early 2000s, this work gave rise to agile, one of today’s foremost team operating systems and a mainstay of complex operations ranging from chip manufacturing to software development.

Rob Tringali is a sports photographer who has captured everything from the Super Bowl and the Olympics to obscure sporting events and high school football games.

At McKinsey, we have created a new team operating system—known as the Way We Work (WWW)—that reframes, reshapes, and improves on the principles of agile teamwork. To do their jobs well, our team leaders need to be able to bring together the best thinking, analysis, coding, and design expertise from our 7,000-plus experts while working with other colleagues to build strong relationships and think strategically. Designed to foster this collaboration across geographies and at scale, WWW can be used by any team, regardless of size, location, or mission.

Consider the example of a Latin American bank that called us in to consult on a transformation of its digital strategy and operations. Things got off to a rocky start. The bank was working not only with us but also with a mix of outside digital architects, designers, and experts. Team members were greatly frustrated: They didn’t know what was expected of them or who was responsible for what. Everyone was working long hours, to little effect. Time to start over.

The team leaders decided to reset using the WWW operating system. First, they held a structured kickoff—involving all team members, including outside consultants—that set a work cadence, team norms, and expectations. Team leaders committed to conducting regular one-on-one feedback sessions and twice-a-month retrospective reviews with the whole team. These systematic take-stock moments, all tracked by a digital tool, enabled the team to make adjustments as needed. During one retrospective, for example, the team realized it needed help applying the bank’s agile methodology to the current transformation project. So it called in experts to provide coaching. By the end of the nine-month project, the bank had successfully launched the digital products it had set out to develop, and the McKinsey team had completed a comprehensive handover, transferring all relevant knowledge to the client. Both the McKinsey team members and the client reported a positive experience—an excellent result after that tough start.

Action plan.

In our research on high-performing teams, we have found that the best teams consistently do the three things that our team at the bank did: hold kickoffs, conduct one-on-ones, and take stock of progress using retrospectives. For our part, we also close out with comprehensive handovers, documenting our work so that the client, and future teams, can build on what we did. We recommend that companies use these practices as the foundation of their teamwork operating system, tailoring them to reflect their important cultural norms. For example, the “obligation to dissent,” one of McKinsey’s core values, is embedded in the design of our one-on-ones and retrospectives.

When we ask our highest-performing teams what they attribute their success to, their answer is: building trust in client relationships. This is what the WWW operating model is meant to facilitate—and we have plenty of evidence that it does. Frequently collecting client feedback, conducting client retrospectives, and holding one-on-ones—our teams listed all these as practices that helped them build trusting client relationships.

As we continue to evolve the WWW operating system, we’ve invested in creating tools to help teams run it. We suggest that companies instituting an operating system do the same. Our teams, for example, can get professional support to facilitate kickoffs and retrospectives, conduct one-on-ones, and set up protocols for handovers. We’ve also created a library of WWW resources that is available to all on our internal website.

[  2  ]

Invest in Active, Real-Time Measurement

To make teamwork scientific, organizations need to be able to measure the outcomes of their actions and determine how changes in the inputs affect results. At McKinsey, we’ve developed tools that allow us to monitor our teams’ needs and performance in real time and then to hypothesize, test, learn, and adapt.

One tool we use to measure and monitor people outcomes is a weekly survey. It is anonymous but segmented, and it is sent to all employees. It asks a single question: How are you feeling? People respond by clicking on one of five emojis, ranging from “Great” to “Terrible.” This simple query allows us to identify trends in sentiment across different colleague groups, such as senior leaders, new hires, data scientists, and editors. If a specific group is slumping, we can intervene—for example, by running targeted programming, offering developmental support, or investing in additional learning and coaching.

To measure functioning, our teams complete an engagement team survey (ETS) every two weeks. It asks members which WWW rituals they are using and how consistently. It also asks how they’re feeling about the project they’re working on and how they’re progressing toward their short- and long-term goals, both personally and professionally. An automated report synthesizes the results. If ETS scores begin to dip, that’s a sign that a team is struggling—and indicates to leaders that it’s time to intervene. To measure how well we’re doing with our clients, we conduct confidential surveys to gather feedback on the quality of our work and the effectiveness of our collaboration.

By combining and evaluating the data from the team and client surveys for a given project, we can identify how and to what extent the operating model used by teams informs our work with clients. For example, we’ve found that teams that use all the elements of the WWW operating system score significantly better on client impact scores than those that do not—from 10 percentage points higher (for collaboration) to 23 percentage points higher (for building client capabilities and skills). Client and team satisfaction often go hand in hand: Teams with the best client feedback scores also have the highest team feedback scores, and vice versa.

Rob Tringali

Internally, we’ve used real-time measurement to figure out whether it makes a difference if our teams work in person or remotely. To answer that question, we studied 40 teams, consisting of 217 consultants, across geographies and sectors, to understand how each model contributed to productivity, client impact, apprenticeship, and personal experience. The results were unequivocal: The sweet spot for in-person time is about 50%. When teams colocate half the time over the course of a project, they’re 10 times as likely to believe they collaborated effectively compared with those above or below that level. Team members are 25% more likely to be satisfied with the quality of their support, and team leaders are significantly more likely to say that they built solid client relationships. For example, 77% of our team leaders said that they had built meaningful connections with their clients when they held problem-solving sessions in person, compared with 59% of those who ran them remotely.

The data also shows that these benefits aren’t linear. Teams that colocate more than 50% of the time increasingly have to make difficult trade-offs when it comes to work-life balance and the ability to conduct focused work. At McKinsey and at other organizations, we’ve found that teams working remotely about 50% of the time are twice as likely to have sufficient recovery from work and are significantly more likely to find their jobs sustainable.

Knowing what to measure is critical. For example, trauma centers determine team effectiveness by measuring how long it takes to stabilize a patient—a major factor in survival. When experimenting with methods to reduce stabilization time, some centers tried placing timers in each trauma bay. Research showed that this simple action significantly reduced stabilization time. The point is that when the right metrics are identified and collected, it becomes clear what practices to implement for the desired outcomes.

In a very different field, but using the same principle, the aviation industry has defined success as fewer crashes—and most important, no fatal crashes. This has been achieved in part owing to new technologies, such as better engines and improved safety systems in the cockpit. But people, too, have been critical in reducing the number of crashes. Carriers, pilots, and regulators have all become a single safety team united by a common goal: no more deaths in the sky. Voluntary incident-reporting programs ensure that those who give information on safety concerns will not be punished for errors. Airlines share data with one another, and pilots are encouraged to recommend solutions. All this data has enabled industry experts to identify best practices, such as revised runway procedures, redesigned flight paths, and better signage. The result: No U.S. commercial carrier has had a fatal crash since 2009, and safety rates for civil aviation have improved too. (See editor’s note below.)

Action plan.

Define what constitutes success. This could be any number of things, such as increased revenue, improved employee and client experience, capability building, work-life sustainability, or skills acquisition. Then build digital systems to track those metrics with as little disruption as possible to teams’ day-to-day work. Finally, use those systems to learn what combinations of inputs lead to better outcomes. At McKinsey, success is delivering results for our clients. The signals that point the way to this outcome are client and team sentiment, and the inputs are the WWW practices, measured through the team survey.

[  3  ]

Create a System for Continuous Improvement and Innovation

The idea of continuous improvement is hardly novel. What is fresh, however, is that teams today have new forms of technology and data collection at their disposal to help them self-correct while projects are underway. At McKinsey, team leaders have access to all the metrics collected and can use them to track progress and assess feedback in real time. On the basis of this data, they can—and are expected to—make improvements. We empower our middle managers to make changes and to run their own experiments. For example, depending on where they are in the course of a project, or depending on changes in team-sentiment data, they might decide to modify their colocation model from being in-person every other week to being in-person for two straight weeks each month.

To facilitate continuous learning, we’ve created a Way We Work center of excellence. This is a centralized group of 14 full-time employees who are responsible for driving the WWW operating system. They also experiment with innovative practices and monitor the relevant flows of data to improve the way teams work. The center can step in if a team is struggling. For example, regional WWW support staff may join a team’s retrospectives to discuss what could have been done better. They also may look at the patterns across teams to identify improvements and share best practices, particularly with regard to the rapid adoption of new technologies, such as generative AI.

Consider what’s happening with the World Economic Forum’s Digital Lighthouse Network. This community of cutting-edge manufacturers applies digital technologies to continuously improve the performance of cross-functional teams. For example, Bosch China incorporates rapid prototyping and testing of new digital tools into standard procedures to help its teams continuously improve speed and quality. Managers at Hindustan Unilever’s Dapada factory use an app to track real-time performance metrics, such as worker productivity, and use the insights to maximize output. In both cases, technology is used to enhance teamwork, not to disrupt it.

Next In

Build Better Teams

Why Leadership Teams Fail

[

And what to do about it

](/2024/09/why-leadership-teams-fail)

At McKinsey, we’ve launched Lilli, a knowledge and capabilities platform that uses generative AI to search and synthesize our entire knowledge library. Lilli is changing the way our teams operate, by accelerating problem-solving and giving team members more time to spend with clients and colleagues.

Action plan.

Identify the metrics that matter most (shift-changeover time, perhaps), hypothesize which actions could improve performance in those areas (preassigned workstations, perhaps), and embed technologies in the operating system (a smart-planning app, perhaps) to enable continuous improvement.

It’s important that teams be able to make their own improvements—something that requires transparency and trust between company leaders and managers. It also requires an operating system that is flexible enough for team members to easily adapt in real time. Continuous improvement can occur only when all perspectives are considered and all teams have access to a centralized knowledge repository.

Finally, it may be useful to set up a center of excellence, staffed with full-time employees with experience in analytics and operating system design. The center’s task would be to identify the rituals, data, and continuous learning practices that are most likely to deliver the best outcomes for a given team.

. . .

Organizations that embrace the three principles we’ve laid out here will find that a feedback loop develops in which everything—from learning to employees to clients to performance is connected. Admittedly, this approach to teamwork requires resources, time, and accountability from senior leaders. But the investment can yield substantial rewards: Our research has shown that executives are five times as productive when working in high-performing teams as they are in average ones. And experience has shown that the most expensive team is one that is struggling.

In a world in which working together is more complex than ever—and more difficult to get right—the new science of teamwork provides organizations with a framework for continuous learning, improvement, and success.

Editor’s note: This article, published in September 2024, references the fact that no U.S. commercial airline carrier has had a fatal crash since 2009. This was true at the time of publication, but on January 29, 2025, an American Airlines flight was involved in a fatal collision with a military helicopter over Washington, DC.