Print Page   |   Your Cart   |   Sign In   |   Register
Diversity in Silicon Valley
Share |


Buck Gee

Denise Peck 

As part of our work researching Asian American demographics in Silicon Valley companies, we scan through countless diversity web pages and annual diversity/corporate social responsibility reports.  All are predictably positive and modestly optimistic while articulating a clear narrative that the company is committed to efforts to improve diversity with programs that are making a difference.    

So it was refreshing to open up Google’s diversity annual report 2018 and read an honest assessment from Danielle Brown, Google’s Chief Diversity and Inclusion Officer, that its diversity and inclusion strategy was not good enough and that changes needed to be made.  Rather than tout its limited “pockets of success” as best efforts, Google exemplified a real corporate commitment to solving its diversity problems by admitting the failures of existing programs and moving to new approaches to the problem.  

We commend Google for the changes in its approach to achieving better diversity outcomes; and we note that the Google’s new approach in the following three areas reflects many of the best practices we would like to see in a strategy to address the Asian glass ceiling. 



“First, the responsibility and work to achieve a more diverse and inclusive Google is shifting from a primarily People Operations and grassroots-led model, to one of shared ownership with Google’s most senior leaders. Google’s leaders are focused on, and committed to accelerating our progress.”


As Brown recognizes, a People Operations (aka H.R.) and grassroots-led diversity effort is unlikely to make significant progress toward diversity goals.  Because we have worked as P/L-focused executives in Silicon Valley and as executive sponsors for diversity initiatives, we have always thought this was obvious.  In our experience, People Operations, without executive support, has little political influence to make major organizational decisions.  Grassroots-led efforts alone, principally led by internal employee resource groups (ERGs), typically have insufficient clout and funding to make a lasting difference.  This is not meant to imply that executives in Silicon Valley do not support diversity.  In fact, all executives we know express support.  However, passive support is not the same as active engagement.  Also, we are not diminishing the importance of an ERG and its role in an inclusive corporate culture, but employee volunteerism is difficult to sustain over time and ill-suited to provide leadership training. 

As one example of executive engagement, Cisco encourages its executives to sponsor at least one extraordinary diverse candidate for career advancement.  As cited in its 2018 Corporate Social Responsibility Report, 42% of Cisco’s executive leadership team and 36% of all vice-presidents have committed to be executive sponsors. 

As another example of active executive engagement, a Bay Area company commissioned an internal task force to examine its internal leadership pipeline several years ago.  In a final presentation to the executive staff, the task force reported little representation in the pipeline for women and all racial minorities.  The business line executives in the meeting argued that they supported diversity, but that the problem would be solved over time as they hired more women and minorities into their overall workforce.  The CEO disagreed and responded by setting out specific objectives, telling his executive team that they would be accountable on meeting those objectives, and empowering his People Operations team to provide leadership development programs for women and minorities. 

We are not sure what Google has in mind specifically with its “shared ownership with Google’s most senior leaders” to accelerate progress, but we would hope that the approach means shared objectives and accountability, working with People Operations to align its diversity programs to the needs of the workforce.  At a minimum, we know that Google has already taken positive steps by upgrading its diversity reporting structure to regularly provide every senior line executive with fine grain reports tracking the impact of diversity efforts in his/her organization.



“Second, we are further increasing transparency. Google’s publication of workforce representation data in 2014 helped shape the current industry conversation on diversity in tech.  We aim to take the conversation—and our work—to the next level as we further refine our approach, so this year we’ve published new and more detailed workforce representation data.” 


Google intends to take the diversity conversation “to the next level” beyond the existing diversity narrative by publishing a more detailed workforce representation data report.  For several years now, Google had already been publishing annual diversity data, making available its EEO-1 reports as a footnoted link in its diversity statements.  Now, it is including much of the EEO-1 information as highlighted topics in the annual diversity report.

We admire Google’s courage and integrity in its commitment to a more honest and open discussion of its diversity challenges and issues hidden in the intersection of gender and race.  In most companies, such information is hidden within People Operations and Legal in fear of fostering internal discontent or external litigation.  We suggest that such information is important for Google employees and its leaders to help uncover insights into the diversity challenges that the company faces and engaging with People Operations to help make Google a more diverse and inclusive workplace.

As one example of a more detailed insight using the report’s intersectional workforce data, consider differences in the representation of white vs Asian women at Google (we use the term “Asian” here to refer to all people of Asian descent as per EEOC definitions).  Using data from Google’s published EEO-1 reports, we present the graph in Figure 1 which charts white and Asian women representation in Google’s overall workforce against their representation in leadership to show that, from 2014 to 2018, white women are a declining share of the workforce while Asian women are an increasing share.  At the same time, white women in leadership has steadily increased while Asian women in leadership has not changed since 2015. 



Does this suggest that Google’s approach to hiring more women is not attracting enough white women?  Or is this due to the higher attrition rates for white women?  Also, does this suggest that Google’s approach to developing women leaders is not developing enough Asian women?

Most companies have the internal information about promotion and attrition rates to answer these questions raised in this example.  We would recommend, as we will argue in the next section of this paper, that companies look at normalized promotion and attrition rates, not only the absolute rates.  Moreover, we would recommend an analysis at each intersection of race and gender and, where justified by the data analysis, consider training programs tailored by specific race and/or gender.  Cisco, for example, developed an internal leadership training program for its Asian American managers.  We understand that Google is developing a program for minority women. 



“Equity: Drive fairness within Google’s processes, as well as in our distribution of resources and opportunity. 
Diversity: Endeavor to attract, develop, progress and retain more underrepresented talent at all levels of Google’s workforce, reaching or exceeding the available talent pool.” 


Earlier this year, we had an “aha”moment in a conversation with the director of diversity and inclusion at one of the largest tech companies in the Bay Area.  We were meeting to describe our “Management Parity Index” analysis of their published EEO-1 reports and how it quantified underrepresentation of Asian Americans in their management and leadership ranks.  He thanked us for our insights, encouraged us to continue our work, and disclosed that their diversity program was primarily driven by compliance to Federal regulations.  He said he would examine what could be done and explained that because Asian representation in their workforce exceeded that in the U.S. population, the company was not obliged to implement talent development or recruiting programs for Asians. 

In other words, the underlying commitment was to regulatory compliance, not workplace diversity. 

Google’s new workforce diversity model, as outlined by Brown, appears to break through the minimal compliance approach with a commitment not just to attract a diverse workforce to Google, but also to equitably develop and progress underrepresented talent at all levels.  We are hopeful that its integrity statement committing to internal corporate values (in addition to and separate from the law) signals a commitment to more than the value of external compliance. 

Specifically, we are hopeful that Google looks beyond the workforce data illustrated in Figure 2, the racial distribution in only one level of its workforce - entry level professional individual contributors in its U.S. white collar workforce (figure charting data from Google’s EEO-1 reports 2013-2017).   

Google’s 2018 diversity report claims modest progress in hiring and retaining black and Hispanic employees in the past few years; and the EEO-1 entry level workforce data shown in Figure 2 confirm more black and Hispanic Googlers.  And as the EEO-1 data in our Figure 2 suggests and the 2018 report confirms, Asians were proportionally the most likely to be hired, of all racial cohorts, by Google in 2017.  41% of all Google hires in the U.S. were Asian, compared to an Asian population of 6% nationally and 20% in the Bay Area. 



As importantly, EEO-1 data suggests that Google is making modest progress in promoting diversity in management levels.  Figure 3 shows that representation for all racial minorities in Google has been consistently increasing since 2013, although Hispanic and black representation as individual contributors and middle managers remain unacceptably low.  On the other hand, Asians appear to be well represented as 29.1% of middle managers. 


However, a more detailed analysis of the data paints a more nuanced picture for each racial cohort, especially for Google’s Asian workforce.  As we argue in our 2017 paper “The Illusion of Asian Success: Scant Progress for Minorities Cracking the Glass Ceiling from 2007-2015”, comparative context is required to properly interpret the data in Figure 3 to understand whether minorities are equitably being promoted into Google management. 

Consider the 2017 representation of Asians (29.1%) in Figure 3. As a measure of upward mobility through the management pipeline, is this good or bad?  As Asians were only 6% of the U.S. population, would 29.1% mean that they were overrepresented? Alternatively, as Asians were 35% of the Bay Area tech workers, would 29.1% mean that they were underrepresented?  

We suggest that neither context is correct if we want to evaluate upward mobility through the management pipeline.  Comparisons with the external U.S. Census population reflect both the success of (i) recruitment from the external population for their employee workforce and (ii) upward mobility of talented employees through the internal pipeline.  A better approach to isolate and measure only upward mobility is to look only at the internal pipeline numbers, comparing the internal executive population directly to the professional population in the pipeline. 

In that context, we use a metric that normalizes the middle management representation to representation as individual contributors.  We call the metric the “Management Parity Index” because, if the representation at both levels are the same, the result is 1.0 and at parity. 



Figure 4 is the MPI analysis for Google’s data derived from Figures 2 and 3.  It suggests that normalized representation in management for blacks and Hispanics have significantly improved since 2013, while normalized representation for Asians has declined.  Representation over time should generally reflect relative promotion rates and, in effect, the data suggests that although Asians have been the most likely to be hired into Google, they have been the least likely to be promoted into management. 



We point out this finding not as a criticism, but to thank Google for  being more transparent in this level of detail and openly admitting the need to refine approaches to improve equity and diversity at all levels.  We are hopeful that normalized analysis of level-to-level promotions by race and gender is a part of that internal refinement. Further internal analysis could help to understand, for a rapidly growing company with a fluid workforce, whether a disproportionately high number of racial minorities were hired in or promoted into management ranks. 


“It is also very important to note that normalized management representation of Asians at Google is above average for Silicon Valley technology companies and higher than 15 of 18 other technology companies that have published EEO-1 reports. Figure 5 is a summary of our Asian MPI analysis of those companies, along with the MPI computed from an aggregation of EEO-1 data (denoted as SV177) from 177 Silicon Valley companies compiled through a unique collaboration of Reveal (from the Center for Investigating Reporting, a nonprofit investigative journalism organization) and the Center for Employment Equity at UMass/Amherst, a research partner of the EEOC with confidential access to EEO-1 datasets.”




Through many formal and informal interactions throughout 2018 with people at Google, we have seen signs that there is an honest understanding that more needs to be done to attract, develop, and promote higher numbers of women and racial minorities, including the development of Asian American talent.  We have attended a community meeting with Danielle Brown, who welcomed direct and honest feedback at the public release of the 2018 diversity report.  We have presented our research findings in discussions with members of its Asian Senior Executive Council (ASEC) and Asian American employee groups in Google.  And we know that Google executives, responding in a uniquely data-driven corporate culture, have become more engaged.

We see that Google is taking positive steps, and we encourage other companies to follow their approach with best practices of proactive executive engagement, more transparency including the public release of its EEO-1 reports, and quest for intersectional equity and diversity at all levels. 

But as Brown herself concludes in the diversity report, “our strategy doesn’t provide all the answers, but we believe it will help us find them.” 

We are hopeful they will. 



Buck Gee is an executive advisor to Ascend. He retired in 2008 from Cisco Systems, where he was vice president and general manager of the Data Center Business Unit. He served as executive sponsor for Cisco's Asian employee resource network and on Cisco's Inclusion and Diversity Council. He joined Cisco with its 2004 acquisition of Andiamo Systems where he was president and CEO. He has also taught computer engineering courses at Stanford University and Howard University. He is a member of the Committee of 100.  He holds BS/EE and MS/EE degrees from Stanford University and an MBA from the Harvard Business School. 


Denise Peck is an executive advisor to Ascend, and is a co-creator and facilitator of its leadership programs for Asian women and millennials. Denise spent fourteen years at Cisco Systems, during which she held vice president positions in marketing, operations, engineering services, and IT, in both San Jose and Shanghai, China. Denise was an executive sponsor and advocate of Cisco’s diversity initiatives, particularly on behalf of women and Asian employee networks. She has a BA degree in economics from U.C. Berkeley and an MBA from the Graduate School of Business at Stanford University. 



The Ascend Foundation ("Ascend") is a 501(c)(3) nonprofit organization providing objective analysis that addresses the challenges facing the corporate community. This paper does not necessarily reflect the opinions of its research clients and sponsors. Ascend disclaims all responsibility and all liability any loss, injury, damage, costs, expenses or compensation of any kind arising directly or indirectly out of or in connection with any act or omission of Ascend in relation to this report including, without limitation, the completeness or accuracy of the information contained in this paper.  

Community Search
Sign In

Latest News
Online Surveys
Featured Members