【Closer Look】Google’s Quest to a High-Performing Team

The Whole is Greater than the Sum of Its Parts

Since 2012, Google has conducted Project Aristotle that aims to discover what makes a high-performance team. The study defines “team” as a highly interdependent entity where they need each other to complete the work. Google emphasizes the difference between “team” and “workgroup” which is characterized by the least amount of interdependency and often based on the organization chart. The concept of team performance is defined by the four dimensions below (Google re:Work, 2016).

1. Executive evaluation of the team
2. Team leader evaluation of the team
3. Team member evaluation of the team
4. Sales performance against quarterly quota

To maintain fairness, the qualitative assessments capture nuances of the results and the culture but have inherent subjectivity. On the other hand, the quantitative metrics provide a concrete measure but lack situational considerations. Thus the four measures in combination enable the researchers to measure how to build a high performing team (Google re:Work, 2016).

Five Factors

With the defined measurements, Google studied 180 teams (115 project teams in engineering and 65 pods in sales) across the company since 2012. With the data collected, the team ran statistical models to understand which of the many inputs would impact the team effectiveness. Using over 35 different statistical models on hundreds of variables, they sought to identify factors that:

1. impacted multiple outcome metrics, both qualitative and quantitative
2. surfaced for different kinds of teams across the organization
3. showed consistent, robust statistical significance

The researchers concluded that how the team worked together mattered more than who is on the team. The factors of contributing to a high performing team are listed below in the order of importance.

1. “Psychological safety: psychological safety refers to an individual’s perception of the consequences of taking an interpersonal risk or a belief that a team is safe for risk-taking in the face of being seen as ignorant, incompetent, negative, or disruptive.
2. Dependability: on dependable teams, members reliably complete quality work on time (vs. the opposite — shirking responsibilities).
3. Structure and clarity: an individual’s understanding of job expectations, the process for fulfilling these expectations, and the consequences of one’s performance are important for team performance.
4. Meaning: finding a sense of purpose in either the work itself or the output is important for team effectiveness.
5. Impact: the results of one’s work, the subjective judgment that your work is making a difference is important for teams. ” (Google re:Work, 2016 )

Doubt

To many researchers’ surprise, the popular notion of team diversity is shown not to have a significant effect on the team performance. This conclusion conflicts with many scholarly articles that emphasize the importance of diversity in team building. Skeptics state that “who” part of the equation doesn’t seem to matter.” In other words, having smart or people from different background does not directly affect how the team performs and thus does not equate to success. It’s a topic that requires further investigation.

In the meantime, it calls attention from researchers that the actual qualitative and quantitative raw data isn’t accessible. Google’s report (named reWork)describes Project Aristotle and links the NYT article but does not provide any portal to the original data by claiming it proprietary. The inaccessibility to the original data makes replication of Google’s results difficult and has raised a question such as how robust the finding is and what is the degree of confidence. A research result is trustworthy to the degree that is it replicable and well-defined, which entails that open data needs to accompany. With complete feasibility of open sourcing the original data, the fact that such source has been omitted calls doubt into Google’s findings.

Motivation and Focus

With Google’s leadership in software technology market, its extensive investment into cracking the code of high team performance, and the scholarly doubt of Project Aristotle, this paper is motivated to examine Google’s finding’s validity.

The focus of this paper is to discover what are the valid factors that influence team performance. Google defined team performance as the indication of whether a work-based team could meet and exceed the target. The five factors proposed by Google are psychological safety, dependability, structure & clarity, meaning and impact.

Psychological Safety First

Google defines psychological safety as “an individual’s perception of the consequences of taking an interpersonal risk or a belief that a team is safe for risk-taking in the face of being seen as ignorant, incompetent, negative, or disruptive.” From practical experience combined with the review of scholarly articles, this concept is the foundation of building a high-performing team.

For evidence, in the 1999 research, “Psychological Safety and Learning Behavior in Work Teams” Dr. Amy Edmondson from Harvard Business school first coined the term psychological safety as “a shared belief that the team is safe for interpersonal risk-taking.” In essence, it resembles the same meaning as Google’s notion. She researched with 51 work-based teams in a company, measuring antecedent, process and outcome variables. Dr. Edmondson undertook qualitative research at a hospital patient-care team. She studies two distinct teams where group A nurses are encouraged to voice out the drug error when spotted — this is an embodiment of established psychological safety in an organization. Group B nurses aren’t encouraged to do so as management team punishes such behavior (Edmondson 1999).

As a result, group A members would openly acknowledge the mistakes and discuss ways to prevent the future occurrence; in contrast, group B members chose to keep the knowledge of the drug error to themselves, thus creating a higher count of total errors in comparison (Edmondson, 1999). To illustrate, a nurse in group A explained matter-of-factly, “mistakes are serious, because of the toxicity of the drugs [we use]-so you’re never afraid to tell the Nurse Manager.” In contrast, a nurse in group B team in the same hospital reported, “You get put on trial! People get blamed for mistakes, you don’t want to have made one.” These quotes illustrate vividly different beliefs about the social consequence. Group B lacks psychological safety in contrast to group A and it’s evident that team information sharing (also known as “speak up”) positively correlated with the extent of psychological safety cultivated within the team. In group A, the open sharing of information helps the team learn about possible reasons for the occurrence of the error and thus could collectively develop solutions to prevent similar mistakes. On the other hand, group B has stifled the information flow within the team, therefore hampered the learning behavior. As a result, with continuous learning enabled by psychological safety, group A delivered a better result than the group B which lacked psychological safety. Dr. Edmondson developed the illustration below to show how a team that encourages psychological safety could transfer its belief into learning behavior and ultimately performance.

Quantitative Evidence

Furthermore, Dr. Edmondson collected extensive data from Office Design Incorporated ( ODI ) a manufacturer of office furniture with approximately 5,000 employees and a reputation for product and management innovation. In total, she had 8 teams that took part in the study to examine the correlation between team performance and the level of psychological safety experienced. Below is the data from her research.

On the generation of all the value above, inspired by Hackman’s team performance scale on obtaining self-report measures of team performance, Dr. Edmondson developed a similar scale for the observer survey, including “this team meets or exceeds its customers’ expectations” and “this team does superb work.” With the highlighted data listed in row 3 team psychological safety and row 6 team performance, it’s evident that as the measure of team psychological safety grows, team performance in turn increases.

A Closer Look

For another research “Making it Safe” completed in 2006 by Dr. Edmondson, she introduced the construct of leader inclusiveness — “words and deeds exhibited by leaders that invite and appreciate others’ contributions. It’s proposed that leader inclusiveness helps cross‐disciplinary teams overcome the inhibiting effects of status differences, allowing members to collaborate in process improvement” (Edmondson 2006). Based on this concept, the research suggests that inclusive leader promotes psychological safety, a key antecedent of engagement in quality improvement work (considered as team performance). Below is the visual illustration of the core argument (Edmondson 2006).

Dr. Edmondson conducted on-site qualitative data collection through 4 of 44 neonatal intensive care units ( NICUs) in the collaborative and interviewed 23 staff members during the visit for 90 minutes sessions. Then the research team conducted a series survey of 1440 health professionals (46% of team members contacted) from 23 NICUs in the United States and Canada completed the survey between July 2003 and May 2004. Of the respondents, 1375 persons declared their profession: 100 as physicians (83 neonatologists, 13 attending physicians, and 4 neonatology fellows), 998 as nurses (65 neonatal nurse practitioners, 16 Master’s prepared nurses, 8 clinical nurses specialists, 867 registered nurses, 14 licensed nurse practitioners, and 28 other nursing functions), 131 as respiratory therapists, and 146 as other types of health care professionals (e.g., social worker, dietitians, etc.). Excluding individuals in the latter group (because the presence and recognition of their discipline as part of the NICU team varied across units), NICUs averaged 60 respondents, ranging from 10 to 164. The average number of respondents per NICU was 3 physician, 15 nurses, and 5 respiratory therapists. From the survey, the positive correlation is shown between the leader inclusiveness, which is the critical driver in cultivating psychological safety and team performance, or in Dr. Edmondson’s term, “engagement in quality improvement work.” Below is the data from her research:

On the generation of the highlighted data above, similarly inspired by Hackman’s team performance scale on obtaining self-report measures of team performance, Dr. Edmondson developed the scale for the observer survey. With the highlighted data listed in row 3 team psychological safety and row 5 team performance, positively correlates with each other (Dr. Edmondson’s term “engagement in quality improvement work”), it’s safe to conclude that team psychological safety plays a vital role in improving team performance.

Then Dependability

Google defines dependability as “on dependable teams, members reliably complete quality work on time.” In essence, it’s the trust that the teammates have for each other in the ability to complete the tasks with quality. For supporting research, Dr. Erdem, Dr. Ozen and Dr. Atsan conducted the research “the relationship between trust and team performance” in 2003. From the research, trust is defined as “team member shave to discover, adopt and perhaps adapt approaches to creating a team spirit in which deep, continual and mutual relationships sustain cooperative behaviors even in the face of extreme pressures” (Erdem, 2003). Fundamentally, it resembles the same meaning as Google’s definition of dependability.

The researchers collected the data, obtained from a sample of ten organizations including 50 teams with 279 members. The selected organizations are the ones which have the most team members in number 28 teams with 148 members). A total of seven teams with 35 members, seven teams with 45 members, seven teams with 31 members and seven teams with 37 members (Erdem, 2003).

With the highlighted data, team performance positively correlates with the sense of trust cultivated within the team. This finding validates Google’s claim that dependability contributes to a high performing team.

More on Dependability

For another supporting case, Dr. Zaheer, Dr. McEvily and Dr. Perrone from Carnegie Mellon University in 1998 conducted the research “does Trust Matter? Exploring the Effects of Interorganizational and Interpersonal Trust on Performance” in 2003. The work defined “interpersonal trust” as the trust placed by individual spanner in his or her opposite member. In other words, it can be interpreted at an interpersonal level, the belief in each other’s ability of reliably completing the tasks, held by all group members. Performance is defined as the work completed on time with quality. The study concluded that interpersonal trust is conducive to the overall team performance.

The research sample is comprised of a list of purchasing managers in the firms who are members of the National Association of Purchasing Managers (NAPM). Extensive semi-structured interviews, every 45 minutes to one hour long, were held with 20 purchasing managers from different firms in the industry. Then the researchers identified 1,050 NAPM members who were eligible to participate in our study by the following criteria

• their firms purchase components
• they deal directly with supplier firms
• they have purchasing relationships with at least six suppliers

Lastly, during the data collection phase, the researchers surveyed 306 purchasing managers and received a total of 205 responses for a final response rate of 67% of individuals eligible and willing to participate (205/306). Of the 205 completed questionnaires received, 120 were from purchasing managers and 85 were from the second respondent in the purchasing organization.

From the highlighted value above, row 3 and row 4 both represent the interpersonal trust level in the surveyed group and the row 9 and 10 represent the performance delivered by the team. The two sets of value are positively correlated and thus supports the claim that interpersonal trust, or dependability (in Google’s term) contributes to a high performing team.

Structure and Clarity

Besides the importance of psychological safety and dependability, a clear team structure should not be overlooked. Team structure and clarity is defined as the extent to which the leader defines team objectives and organizes the team to achieve them as well as the consequences of one’s performance are essential for the team effectiveness (Google, 2016). The common practice for goal setting at Google is through Objective and Key Results (OKRs) to help communicate short and long-term goals, both quantitatively and qualitatively.

For supporting research, Dr. Michael West’s study in 2003 “leadership clarity and team innovation in health care” revealed that how the “clarity of and commitment to team objectives” could yield in high team performance. The measurements are defined below.

• The measure of “clarity and commitment to team objective” is “measured using 11 items to assess the extent to which team members are clear about their work-related objectives and the extent to which they perceive objectives to be worthwhile. Using a seven-point response scale ranging from not at all to complete, respondents answered questions about their team’s objectives. One example is ‘‘How clear are you about your team’s objectives?’’
• The measure of “team innovation” is “ team members were asked to write descriptions of the major changes or innovations the team had introduced in the previous 12 months. Team innovations were rated using a five-point scale on four dimensions (West & Anderson, 1996): magnitude — how great would be the consequences of changes introduced and to what extent changes would improve the team effectiveness.”

Dr. West’s defined team innovation is “to what extent changes would improve the team effectiveness.” The term team effectiveness defined here, by nature is the same as Google’s concept of team performance (West 2003).

Evidence

Dr. West collected a sample that comprised 3447 respondents from 98 primary health care teams (PHCTs), 113 community mental health teams (CMHTs), and 72 breast cancer care teams (BCTs). The results revealed how team process (its dominant factor is clarity and commitment to team objective) could positively yield in (team) innovation. Below is the data from his research.

From the table above, the highlighted value illustrate how (team) innovation is positively correlated with team process (its dominant factor is clarity and commitment to team objective, explained in West’s paper) with statistical significance. In essence, Dr. West’s study supports Google’s claim that team structure and clarity contribute to building a high performing team.

Analysis of Structure and Clarity

For another supporting research, “an examination of goal and process clarity” in 2011 done by Dr. Hu, Jia and Dr. Liden, Robert C.. from Department of Managerial Studies, University of Illinois, Chicago. This study investigated goal and process clarity as an antecedent of team potency and subsequent team effectiveness.

• The authors defined the goal and process clarity as “one needs to have clear expectations about (a) one’s subgoals, (b) the paths to accomplish these subgoals, and © the link between one’s work and the work of others.” This essentially is the same as how “structure and clarity” is defined by Google.
• This definition of Team potency, defined as team members’ shared beliefs about their collective capabilities is a critical motivational state in teams.

Dr. Hu and Dr. Liden’s believe that “team potency serves as a bridge linking goal and process clarity to team effectiveness, such that goal and process clarity are positively related to team potency, which in turn leads to team effectiveness” as illustrated below. After carefully reviewing the research, it’s notable here the meaning of “team effectiveness” can be used interchangeably with Google’s notion of team performance (West & Linden 2011) .

Evidence from China

Dr. Hu and Dr. Linden collected data from five banks in China, where the employee worked in different teams such accounting, operation and management. Each team developed a common team goal and every member held onto a clear role and worked effectively towards the team objective. To ensure the full team member, Dr. Lu and Dr. Liden only included employees who had been with the team over 6 months. There were 570 employees, on behalf of 95 teams that completed surveys. Seven teams’ surveys were discarded due to the lack of senior manager ratings. Thus a total of 71 teams was included in the analyses, with 304 employees (response rate 53.3%) and 60 upper-level managers (response rate 75%) forming the final sample (Hu & Linden, 2003). Below is the data from the research.

As the highlighted value shown above, it confirms the authors hypothesis that goal clarity and process clarity have a total positive effect on team performance with statistical significance. In further discussion, it’s evident and intuitive that when team members have a good understanding of the objectives and the procedure between their tasks linked to the common objective. This is motivating the teammates to translate their endeavor into achieving common team objective.

Impact

Google claimed that “impact” played an important role in building a high performing team. The idea of “impact” is defined as results of one’s work, the subjective judgment that your work is making a difference is important for teams. In practice, team leader often uses external rewards to help shape the teammates’ subjective opinions that their work matters.

For supporting research, Dr. Sarin and Dr. Mahajan conducted an extensive study “The effect of Reward Structures on the Performance of Cross-Functional Product Development Teams” in 2001. The study reveals that for long and complex projects, outcome-based rewards have a positive effect on overall team performance (Sarin & Mahajan, 2001). Dr. Sarin and Dr. Mahajan define outcome-based reward as the degree to which team rewards are tied to the bottom line/profitability of the product development by the team. Outcome based reward is the effective way to help members validate that their contribution matter to the organization or “impactful” (in Google’s term).

Besides, team performance is measured based on the five specific dimensions below

• Speed to Market
• Adherence to Budget and Schedule
• Level of Innovation
• Product Quality
• Market Performance

Dr. Sarin and Dr. Mahajan conducted 57 qualitative field interviews from 26 teams and surveyed 246 members of 65 teams, drawn from 13 divisions of 6 medium-sized to large organizations. The teams varied in size from 3 to 22 members, with an average of 7.8. By definition, these teams were temporary, having worked together for anywhere from 3 to 72 months. The result of the data analysis is below:

As the highlighted value above, it confirms the researchers that outcome-based reward has a positive effect on the overall team performance across the five dimensions. In further discussion, the external reward ensures the teammates that their contribution is valued by the organization. Thus by Google’s definition, teammates feel a strong sense of impact would deliver high-performance.

Final Word

Building high performing team is incredibly valuable in the modern business landscape. Google’s attempt to crack the team high-performance code is inspiring to the industry. Google proposed five key concepts that built a high performing team: 1, psychological safety 2, dependability 3, structure and clarity 4, meaning and 5, impact. This paper serves to examine its five factors with other scholarly research and articles. The analysis above validated the four factors, including psychological safety, dependability, structure & clarity and impact. However, this research has not found meaningful qualitative and quantitative evidence to support how “meaning” could contribute to building a high performing team. It will require further research, assuming Google’s would release the original sample data. Overall, the pursuit of team excellence is a valuable topic as that’s the fuel of the innovation for humanity and Google has taken an applaudable leap.

Thanks

Akbar Zaheer, Bill McEvily, Vincenzo Perrone, (1998) Does Trust Matter? Exploring the Effects of Interorganizational and Interpersonal Trust on Performance. Organization Science 9(2):141–159. https://doi.org/10.1287/orsc.9.2.141

Edmondson, A. (1999). Psychological Safety and Learning Behavior in Work Teams. Administrative Science Quarterly, 44(2), 350–383. doi:10.2307/2666999

Ferda Erdem, Janset Ozen, Nuray Atsan, (2003) “The relationship between trust and team performance”, Work Study, Vol. 52 Issue: 7, pp.337–340,https://doi.org/10.1108/00438020310502633

Google. “Re:Work.” Google, 1 (Mar. 2016) rework.withgoogle.com/print/guides/5721312655835136/.

Hu, J., & Liden, R. C. (2011). Antecedents of team potency and team effectiveness: An examination of goal and process clarity and servant leadership. Journal Of Applied Psychology, 96(4), 851–862. doi:10.1037/a0022465

Michael A West, Carol S Borrill, Jeremy F Dawson, Felix Brodbeck, David A Shapiro, Bob Haward (2003) Leadership clarity and team innovation in health care,The Leadership Quarterly, Volume 14, Issues 4–5, 2003, Pages 393–410, ISSN 1048–9843, https://doi.org/10.1016/S1048-9843(03)00044-4

Nembhard, I. M. and Edmondson, A. C. (2006), Making it safe: the effects of leader inclusiveness and professional status on psychological safety and improvement efforts in health care teams. J. Organiz. Behav., 27: 941–966. doi:10.1002/job.413

Shikhar Sarin, Vijay Mahajan (2001) The Effect of Reward Structures on the Performance of Cross-Functional Product Development Teams. Journal of Marketing: April 2001, Vol. 65, №2, pp. 35–53. https://doi.org/10.1509/jmkg.65.2.35.18252

***Special thanks to Prof. Buck Goldstein for his incredible mentorship on creating this writing.