JANUARY 2016

Revisiting the matrix organization

Michael Bazigos and James Harter

Matrices are often necessary, but they may create uncomfortable ambiguity for employees. Clarifying roles can boost both the engagement of the workforce and a company’s organizational health.

Matrix organizations have been around Health Index (OHI), clear and account- for decades, stimulating vigorous able roles are among the most important debate between supporters and detrac- drivers of organizational health. Taken tors for nearly as long.1 They remain together, the Gallup and McKinsey prevalent at the large number of !ndings underscore how important it is companies that need to bring functional for executives and line managers to centers of excellence together with address the role ambiguity that’s all too business-speci!c people and processes. common in matrix organizations. Eighty-four percent of respondents (For more on the research behind these to a recent Gallup survey, for example, two studies, see sidebar, “About were at least slightly matrixed. the research,” on page 4.)

That survey, covering nearly 4,000 workers in the United States, highlights Ubiquitous and unexceptional some bene!ts for employees in matrices, particularly in areas related to Eighty-four percent of the US employees collaboration. At the same time, the Gallup surveyed were matrixed to survey suggests that these employees some extent. Forty-nine percent served feel less clear about what’s expected on multiple teams some days (we of them than their nonmatrixed counter- categorized them as slightly matrixed), parts do. This problem has conse- and 18 percent served on multiple quences: Gallup research indicates that teams every workday but with different clarity of expectations is a foundation people, though mostly reporting to the for building an engaged workplace that same manager (matrixed). The remaining performs at high levels. Furthermore, 17 percent reported to different man- according to McKinsey’s Organizational agers in their work with different teams (supermatrixed). 2

Most employees in matrixed organiza- were generally about twice as likely as tions, according to the survey, aren’t slightly matrixed ones to say that terribly engaged with their jobs. (Gallup their organizations not only helped them de!nes employee engagement as collaborate more effectively with involvement in and enthusiasm for work.) coworkers, do their best work, and These !gures are consistent with what serve customers well but also stimulated Gallup has found in the workplace at bottom-up innovation. Supermatrixed large over a decade of study. They are employees were also somewhat more alarming, given the relationship between likely than those in the other categories worker engagement and vital business to say they had received recognition outcomes, such as productivity, pro!t- or praise during the past seven days, that ability, and customer perceptions of their opinions counted, and that their service quality.2 The survey does sug- fellow employees were committed to gest a modestly positive relationship doing quality work. These are key between the four categories of organi- elements in the overall engagement of zation and employee engagement, employees and suggest that relationships which rises slightly across them (exhibit). and collaboration among employees in matrixed organizations and their peers and superiors really are better. Collaboration and clarity On the other hand, only a minority of Beneath the surface, we found some the supermatrixed employees strongly areas (particularly collaboration) where agreed with the statement, “I know matrixed organizations performed what is expected of me at work,” com- better than less matrixed ones and others pared with 60 percent of the nonmatrixed. (related to role clarity) where they did This re"ects a common complaint worse. The differences in engagement about matrixed organizations—that the at more and less matrixed organizations structure gives rise to a lack of clarity suggest advantages and disadvantages about responsibilities, expectations, and that may cancel one another out. who reports to whom. Workers in the three matrixed groups were more likely A key area of strength for matrixed than nonmatrixed ones to say that organizations lies in collaboration—a they need clear direction from project heartening discovery, since cross- leaders and communication between company teamwork is one of the chief their managers and project leaders to aims of many matrices. We asked prioritize their work most effectively. employees of slightly matrixed, matrixed, and supermatrixed organizations Also, employees in the matrixed cate- about the bene!ts of being on different gories were more likely than their teams. Supermatrixed employees nonmatrixed counterparts to say they 3

Exhibit Matrixed employees are slightly more engaged.

% of US employees1 Engaged Not engaged Actively disengaged

Nonmatrixed 28 56 16

Slightly matrixed Work on multiple teams 29 57 14 on some days

Matrixed

Work on multiple teams 31 55 14 every day, primarily reporting to same manager

Supermatrixed2 Work on multiple teams 34 56 11 every day, reporting to different managers

1Controlled for employment level. Data re!ect merged responses from 2 surveys and are not weighted. 2Figures do not sum to 100%, because of rounding.

Source: Gallup

spent their days responding to The link to organizational health coworkers’ requests and attending internal meetings. Such responses are Interestingly, role clarity and related not surprising in an environment accountability practices emerge as where employees receive instructions among the most important drivers of and feedback from multiple managers organizational health, and ultimately and work with a range of people to com- performance, in McKinsey research plete projects. These are also probably based on the Organizational Health factors in the critics’ assertions that the Index (OHI). McKinsey has consistently matrix structure can slow decision found that improving role clarity improves making, blur lines of communication, accountability, an outcome that is a sti"e productivity, and hinder organiza- critical component of the overall health- tional responsiveness and agility.3 index score. In fact, organizations with 4

About the research

The !ndings of the study on matrixed based on three capabilities: aligning around employees are based on a Gallup panel strategies, executing them, and adapting web survey, completed by 3,956 full-time when necessary.1 The index includes data employees aged 18 and older, that was from more than two million respondents and administered between April 8 and April 27, over 2,000 unique surveys. Organizations 2015. The Gallup panel is a probability- in the top quartile for health collectively based longitudinal group of US adults outpace organizations in the bottom quar- selected through random-digit-dial (RDD) tile in total returns to shareholders (TRS): phone interviews over landlines and cell they earned three times the annual TRS of phones. Address-based sampling methods bottom-quartile organizations over the are also used to recruit panel members. nine-year period of the study.2 The Gallup panel is not an opt-in panel, and members are not given incentives The study focusing on the accountability for participating. practices of organizations was conducted using data from 254 unique companies Our sample for this study, which used and 781,224 respondents, collected in 2014 Current Population Survey !gures, was and 2015. This study determined the weighted to be demographically repre- rank order of practices structurally related sentative of the US adult population. For to organizational-health outcomes. The results based on this sample, the maxi- order of the practices was based on the mum margin of sampling error is plus or magnitude and signi!cance of the minus two percentage points at the standardized betas produced by regressing 95 percent con!dence level. Margins of the outcome on the direct practices. error are higher for subsamples. In addition To determine the rank order of the related to sampling error, the wording of ques- practices, we !rst regressed the outcome tions and practical dif!culties in conducting on the direct practices and then (using surveys can introduce error and bias a stepwise regression) entered the remain- into the !ndings of public-opinion polls. ing practices. Practices that explained The survey responses were matched a minimum incremental 1 percent of the with those of a US workforce panel survey variance were labeled related practices. administered in November 2014 to study Their rank order (like our treatment of direct the engagement and other work-related practices) was based on the incremental factors of matrixed employees. Separately, amount of variance explained. Gallup’s meta-analysis of the relation- 1 ship between employee engagement and Organizational health is operationally defined by scores on nine organizational outcomes: direction, business outcomes included more than leadership, culture and climate, accountability, 49,000 business units across 49 industries. capabilities, coordination and control, innovation and learning, motivation, and external orientation. Unlike employee engagement, they are assessed The results of the organizational studies by survey questions about the organization’s are based on subsets of McKinsey’s global effectiveness in these areas rather than their impact database for the Organizational Health on employees. Index (OHI). This index is a survey-based 2 See Aaron De Smet, Bill Schaninger, and Matthew assessment of organizational health, Smith, “The hidden value of organizational health— and how to capture it,” McKinsey Quarterly, April de!ned as the ability to perform over the 2014, on mckinsey.com. long term. That kind of performance is 5

high accountability scores have a Priorities for matrixed managers 76 percent probability of achieving top- quartile organizational health—more than Given the importance of role clarity and triple the expected rate. What’s more, accountability to organizational health the independent effects of role clarity are and, ultimately, performance, address- so powerful that they affect OHI scores ing the role ambiguity that pervades directly, one of only four management matrixed companies is a critical priority practices (among 37) that do.4 for their leaders, who should help employees by continually setting clear These !ndings are consistent with work expectations aligned with the direc- by McKinsey’s Suzanne Heywood and tion of the business. This clarity should others showing that organizations can cascade into frequent conversa- mitigate the complexity associated tions between managers and their direct with matrices through clear accountability reports about the speci!c role each and targets for individuals.5 Further person plays in advancing the company’s reinforcing these !ndings is the academic objectives. Consultative (as opposed literature suggesting that higher levels to authoritarian) leadership practices can of the ownership mentality predict higher contribute meaningfully to accountability, levels of collaboration, organizational according to McKinsey’s OHI research. commitment, and corporate citizenship, as well as reduced levels of behavior It is also imperative to maintain day- that deviate from workplace norms.6 to-day lines of communication to root out and dispel ambiguity and ensure The Gallup survey does suggest that that everyone is consistently on the role clarity takes a hit in matrixed organi- same page. This is true at the organiza- zations. Yet it also indicates that super- tional as well as the team level: Gallup matrixed employees were more likely to research shows that managers should have received recognition or praise not save critical conversations for in the previous seven days and to believe once-a-year performance reviews— that their opinions counted. McKinsey engagement "ourishes when employees research suggests that these features of receive regular, actionable feedback on the employee experience in matrixed their progress. companies have a positive impact on organizational health: two management Last, the matrix structure is notorious for practices—recognition and employee frequently obscuring lines of account- involvement in direction setting—are ability, so leaders and managers should important drivers of two of the OHI’s ensure that all employees understand outcomes—motivation and direction— whom they answer to and the duties for which, along with accountability, are which they are responsible. The impor- meaningful components of the overall tance of regular discussions to reclarify OHI score. expectations as work demands change 6

is compounded in matrix organizations. 5 See Suzanne Heywood, Jessica Spungin, and And highly engaged employees thrive in David Turnbull, “Cracking the complexity code,” McKinsey Quarterly, 2007 Number 2; a system where everyone is account- and Suzanne Heywood and Julian Birkinshaw, able for his or her work. “Putting organizational complexity in its place,” McKinsey Quarterly, May 2010, on mckinsey.com.

6 James B. Avey, Bruce J. Avolio, Craig D. Crossley, 1 For a classic critique, see Tom Peters, “Beyond and Fred Luthans, “Psychological ownership: the matrix organization,” McKinsey Quarterly, Theoretical extensions, measurement, and September 1979, on mckinsey.com. relation to work outcomes,” Journal of 2 James K. Harter et al., The relationship between Organizational Behavior, 2009, Volume 30, engagement at work and organizational Number 2, pp. 173–91; doi 10.1002/job.583. outcomes, Gallup, February 2013, gallup.com. 3 For more, see Billie Nordmeyer, “Disadvantages The authors wish to thank Gallup’s Sangeeta of organizations with the matrix approach,” Agrawal, Annamarie Mann, and Susan Houston Chronicle, chron.com; and Eric Krell, Sorenson, as well as McKinsey’s Lili Duan, “Managing the matrix,” HR Magazine, Society for Human Resource Management, April 1, 2011, Dominik Deja, Dinora Fitzgerald, and Yuan shrm.org. Tian, for their contributions to this article. 4 Since the impact of these practices transcends geography, industry sector, and company size, Michael Bazigos, head of organizational we call them power practices. Besides role clarity, science at McKinsey, is based in McKinsey’s personal ownership (another accountability practice), strategic clarity, and competitive New York of!ce. Jim Harter is the chief insights are also in this select group. Overall, we scientist of workplace management and well- assessed the 37 management practices through being for Gallup’s Workplace Management empirically derived survey items that were independent of the outcomes they predicted. We Practice in Omaha, Nebraska. assessed the independent effect of role clarity after statistically controlling for shared or Copyright © 2015 McKinsey & Company. overlapping effects among the 37 practices. All rights reserved.