Moussouris v. Microsoft Corp.

Full title: KATHERINE MOUSSOURIS, et al., Plaintiffs, v. MICROSOFT CORPORATION…

Court: UNITED STATES DISTRICT COURT WESTERN DISTRICT OF WASHINGTON AT SEATTLE

Date published: Jun 25, 2018

CASE NO. C15-1483JLR (W.D. Wash. Jun. 25, 2018)

Opinion

CASE NO. C15-1483JLR

06-25-2018

KATHERINE MOUSSOURIS, et al., Plaintiffs, v. MICROSOFT CORPORATION, Defendant.

JAMES L. ROBART United States District Judge

ORDER DENYING CLASS CERTIFICATION

I. INTRODUCTION

Before the court is Plaintiffs Katherine Moussouris and Holly Muenchow's (collectively, "Plaintiffs") motion for class certification. (MCC (Dkt. ## 228 (sealed), 381 (redacted)).) Defendant Microsoft Corporation ("Microsoft") opposes the motion. (Resp. (Dkt. # 474).) Microsoft also filed a surreply to strike portions of Plaintiffs' reply. (Surreply (Dkt. # 349).) The court has considered the motion, the submissions filed in support of and in opposition to the motion, the relevant portions of the record, and the applicable law. The court also heard oral argument from the parties on June 11, 2018. (See Min. Entry (Dkt. # 503).) Being fully advised, the court DENIES the motion for the reasons detailed below.

II. BACKGROUND

Microsoft is a "global provider of software and software-related services as well as hardware devices." (SAC (Dkt. # 55) ¶ 2.) Microsoft recognizes that diversity "is an integral and inherent part of [its] culture, fueling [its] business growth while allowing [it] to attract, develop, and retain [the] best talent." (Parris Decl. (Dkt. ## 288 (sealed), 287 (redacted)) ¶ 3, Ex. 1 ("Resp. Docs.") at 354.)¹ To that end, Microsoft devotes significant resources to various diversity and inclusion ("D&I") initiatives, including D&I training courses, D&I plans to recruit and retain diverse employees, and D&I toolkits to aid internal discussions regarding diversity. (Dhillon Decl. (Dkt. ## 297 (sealed), 387 (redacted)) ¶¶ 6-13.) Microsoft's official policy ("Anti-Harassment/Anti-Discrimination policy") prohibits all workplace discrimination and expresses zero tolerance for any form of workplace harassment. (Resp. Docs. at 18, 173.) Microsoft's Employment Relations Investigation Team ("ERIT") investigates any alleged violations of that policy. (Parris Decl. ¶ 7, Ex. 5 ("DeLanoy Dep.") at 31:3-8.)

1 For the set of Microsoft corporate documents filed with the responsive briefing, the court cites the page number located at the bottom center of the exhibit. (See generally Resp. Docs.)

Plaintiffs—former and current Microsoft employees (SAC ¶¶ 6, 8)—challenge Microsoft's "continuing policy, pattern, and practice of sex discrimination against female // // employees in technical and engineering roles . . . with respect to performance evaluations, pay, promotions, and other terms and conditions of employment." (Id. ¶ 1.) Plaintiffs claim that, as a result of Microsoft's employment policies and practices, female technical employees "receive less compensation and are promoted less frequently than their male counterparts." (Id. ¶ 3; see also id. ¶ 25 ("Microsoft discriminates against female technical employees in (1) performance evaluations; (2) compensation; and (3) promotions.").) Plaintiffs now seek to certify a proposed class of female employees in Stock Levels 59-67 working in the Engineering and/or the I/T Operations Professions from September 16, 2012, to the present. (MCC at 1.)

The court details Microsoft's structure, relevant employment practices, and history with gender equity issues before summarizing Plaintiffs' employment and the relevant procedural background.

A. Microsoft's Structure

Microsoft categorizes its employees into various groups. First, Microsoft groups its employees into Professions, the "highest level in Microsoft's taxonomy" (Whittinghill Decl. (Dkt. ## 321 (sealed), 320 (redacted)) ¶ 7), defined as "a group of functional areas . . . with common functional skillsets, business results, and success differentiators," (Shaver Decl. (Dkt. ## 229 (sealed), 233 (redacted)) ¶ 5, Ex. A ("MCC Docs.") at 757).² Within Professions, Microsoft classifies employees into Disciplines, subgroups that // represent "a different area of focus within a Profession." (Whittinghill Decl. ¶ 8.) Microsoft further breaks down Disciplines into Standard Titles, which "represent a different role within a Discipline." (Id. ¶ 10.) For example, an employee can be in the Engineering Profession in the Program Management Discipline with the Standard Title of Program Manager. (Shaver Decl. ¶ 9, Ex. E ("Whittinghill Dep.") at 103:20-104:1.)

2 For the set of Microsoft corporate documents filed with Plaintiffs' motion, the court cites the ECF-generated page number at the upper right-hand corner of the exhibit. (See generally MCC Docs.)

Microsoft further divides Standard Titles into areas of specialization. (See Whittinghill Decl. ¶ 21.) This further subdivision, called a "functional hierarchy," features six tiers of division. (See id. ¶ 22.) Oftentimes, this functional hierarchy—rather than the larger groupings of Profession, Discipline, or Standard Title—tracks reorganizations in the business that may affect an employee's responsibilities. (Id.) Thus, an employee may "keep the same Standard Title, Career Stage, and Stock Level, but end up working on very different products and/or services." (Id.)

The functional hierarchy reveals that there may be numerous subdivisions within a Standard Title. (See id. ¶¶ 34-35.) For example, within the Data & Applied Scientist Standard Title, the functional hierarchy describes 268 unique positions that employees hold. (Id. ¶ 35, Ex. C.) As a result, job positions with the same Standard Title may describe significantly different roles. (See id. ¶¶ 39-40, Ex. D (comparing a "Software Engineer II" on the Azure, Internet of Things team to a "Software Engineer II" on the Xbox team).) The number of unique positions varies from year to year because Microsoft alters its team structure in response to market demands. (Id. ¶ 37.) Thus, in any given year, Microsoft is likely to form new teams or eliminate old teams. (See id.) Those changes can generate or remove employee roles. (See id.)

Microsoft also utilizes Stock Level and Career Stage classifications. Stock Level, also known as "Pay Level," represents "compensation ranges" that are set "relative to the professions [Microsoft] employ[s] and the markets in which [Microsoft] work[s]." (MCC Docs. at 760; see id. ("Pay levels (such as levels 50, 59, or 64) represent salary ranges, based on an ongoing analys[i]s of local and discipline-specific labor markets and Microsoft's compensation strategy.").) Stock Levels range from 59 to 98 (see Shaver Decl. ¶ 10, Ex. F ("Ritchie Dep.") at 521:10-13; Farber Rep. (Dkt. ## 332 (sealed), 384 (redacted)) ¶ 20), and pay increases as the Stock Level increases (see Farber Rep. ¶ 20; Saad Rep. (Dkt. ## 354-1 (sealed); 385 (redacted)) ¶ 145). Career Stage "indicates at a high level the general degree of scope and impact of a role." (Whittinghill Decl. ¶ 15.) There are three Career Stage designations. An "Individual Contributor" ("IC") is "an entry-level employee with little or no relevant experience." (Id. ¶ 16.) A "Lead" is "a professional who primarily contributes as an IC but also supervises a small team of ICs." (Id. ¶ 17.) A "Manager" "manag[es] ICs and/or other managers." (Id. ¶ 18.)

B. Microsoft's Pay and Promotion Processes

Plaintiffs challenge a specific aspect of Microsoft's employee evaluation process: the Calibration Process, during which managers compare and standardize employees' performance ratings across a cohort of similarly-leveled colleagues. (See Ritchie Dep. at 136:24-137:4, 157:6-8; see also Resp. Docs. at 55; DeCaprio Decl. (Dkt. # 295) ¶ 4; Helf Decl. (Dkt. ## 301 (sealed), 300 (redacted)) ¶ 4.) The Calibration Process took two forms over the course of the class period. From 2011 to 2013, Microsoft called the comparison processes "calibration meetings," with "a focus on the results achieved by the employee and how he or she achieved those results." (Wilson Decl. (Dkt. # 322) ¶ 9.) In May of 2014, Microsoft altered components of this process and renamed calibration meetings "people discussions." (DeCaprio Decl. ¶ 4.) The court discusses "calibration meetings" and "people discussions" before addressing the evaluation of the Calibration Process conducted by Plaintiffs' expert Dr. Ann Marie Ryan.

1. Calibration Meetings

Managers held calibration meetings to discuss employees' performance ratings relative to their similarly-positioned peers to determine the final performance ratings for each employee. (Resp. Docs. at 55; Helf Decl. ¶ 4.) Managers held those meetings at the end of the assessment process, after they had reviewed self-evaluations, feedback from others, and their own observations to recommend an initial Performance Rating from 1 to 5, with 5 as the lowest and 1 as the highest. (See Resp. Docs. at 39; Helf Decl. ¶ 9.) Managers based their initial ratings on three inputs: "What results were achieved, How they were achieved, and [the] Proven Capability" of the employee over time. (Resp. Docs. at 39 (bolding in original).) To aid managers' evaluation of the "what" and the "how," Microsoft provided general definitions and a rubric describing what each numerical rating represented. (Id. at 50-51.) For instance, Microsoft defined the "What" as evaluating the employee's "[r]esults against commitments for the past fiscal year." (Id. at 50.) Additionally, Microsoft noted that a "1" should be given out to approximately 20% of employees who "[g]reatly surpassed expectations with performance and business impact greater than the vast majority of peers." (Id. at 51.) //

Managers then handed over their initial performance ratings to and discussed them with the Calibration Manager, who represented a number of employees at a calibration meeting. (Id. at 52-53.) At the calibration meeting, Calibration Managers explained the reasoning for employees' initial performance rating and compared the relative performance of employees doing similar work. (DeCaprio Decl. ¶ 4.) Employees were grouped by Stock Level into peer groups for these comparisons. (MCC Docs. at 13.) The goal of the meeting was to "ensure there was consistency to how performance was evaluated and rated vis-a-vis the performance of [an employee's] peers." (Wilson Decl. ¶ 9.) Microsoft encouraged managers to refrain from making "artificial distinctions . . . simply to meet an approximate distribution" of how many employees should receive a certain rating. (Resp. Docs. at 55.) Thus, Microsoft recognized that an "extraordinary high-performing group may . . . justify a distribution that is skewed to the higher end of an approximate distribution," whereas an "underperforming group may . . . justify a distribution that is skewed to the lower end." (Id.)

Beyond this guidance, Microsoft sought to "empower leaders and managers to use their judgment in evaluating the performance of their employees" by localizing the evaluation process to each group. (Wilson Decl. ¶ 11; see also Helf Decl. ¶ 6 (stating that the calibration meetings "looked different depending on the level of the roles for the employees involved and . . . the way the leader wanted to run the meeting").) Microsoft encouraged every meeting leader to create the approach to be used in his or her calibration meeting, including what issues and topics to discuss; how the meeting would be conducted; and the specific qualities that each group would look for in evaluating employees. (Wilson Decl. ¶ 13; id. ¶ 21 (describing calibration meetings as "designed to be flexible").) Thus, although Calibration Managers ran the meetings according to Microsoft's framework, "the managers participating in the calibration meeting" determined "the common core priorities . . . for the larger group," any "common objectives," and "what good work looked like." (DeCaprio Decl. ¶ 8.)

As a result, "the personalities in the room" largely dictated how calibration meetings unfolded. (Wilson Decl. ¶ 15; see also Helf Decl. ¶ 15 ("No two [c]alibration [m]eetings that I attended were the same.").) Calibration meetings thus varied "both structurally and substantively." (DeCaprio Decl. ¶ 9.) Structurally, the meeting leader determined how the discussion would be held, such as through the use of a PowerPoint versus the use of physical cards to sort employees. (Id.) Substantively, the discussions focused on different issues, such as budgetary needs versus employee talent. (Id. ¶ 10.)

Managers similarly discussed promotions during calibration meetings. In determining promotions, managers considered business need, demonstrated employee readiness, and available budget. (Helf Decl. ¶ 30; Resp. Docs. at 38.) Microsoft, as with the performance ratings, provided general definitions of the three promotion criteria. (See Resp. Docs. at 38.) Because promotions depended upon the budget in any given year, promotions varied across levels, locations, Professions, and year. (See id. at 36.)

The calibration meeting recommendations were provided, or "rolled up," to the next level of management for review. (See Wilson Decl. ¶ 24.) In other words, the recommendations reached by lower-level managers "cascaded upwards" for approval by higher-level managers. (See Helf Decl. ¶ 21.) Higher-level managers did not usually make significant adjustments to the individual recommendations but would instead "focus[] on the budget and ensur[e] the actual distribution was in line with the budget." (Id. ¶ 27.) The roll up process continued until the recommendations reached the Executive Vice President ("EVP"), who "in practice . . . did not change recommendations made at" lower-levels. (Id.; see also Wilson Decl. ¶ 24.)

2. People Discussions

Since 2014, Microsoft "moved away from lengthy meetings for comparing and evaluating employees" (Helf Decl. ¶ 31) and instead shifted its evaluation process to focus on "the impact the employee's performance makes on his or her organization and across the company in light of his or her role" (Wilson Decl. ¶ 10). In this new evaluation process, managers held "people discussions" to evaluate the impact of an employee's performance. (Id.) Unlike before, Microsoft no longer had "numerical performance ratings or a company-wide anticipated distribution of performance ratings or annual rewards." (Id.; see also DeCaprio Decl. ¶ 4.) Thus, the "people discussion" meeting often focused on "what rewards are right for each individual employee based on the impact that particular employee's performance had on the group and the company." (Wilson Decl. ¶ 23; see also Helf Decl. ¶ 4.)

Aside from the change in focus, the people discussions resembled the calibration meetings. Like calibration meetings, the process rolled upwards, with higher-level managers reviewing the recommendations made by lower-level managers. (See Helf Decl. ¶ 36.) And as was the case previously, higher-level managers are less likely to substantively change recommendations, especially those concerning lower-level employees. (See id.) This roll-up process continued until the head of the organization approved the recommendations. (See id.; Wilson Decl. ¶ 24.)

The people discussion process "empower[ed] managers to own the assessment by focusing solely on the individual employee instead of on the individual as compared to others." (Helf Decl. ¶ 31.) Thus, Microsoft "encourage[d] leaders to take more direct ownership of the evaluation process." (Id. ¶ 35.) Managers "ha[d] much more latitude to determine the priorities for their organization and how to accomplish those priorities," including what inputs to consider during the people discussions and which employees to discuss. (Id.; DeCaprio Decl. ¶ 14 (observing variations in people discussions, such as who was discussed and how those discussions were run).) Managers could even define for themselves "what high impact for different types of roles looks like, such as what type of impact do [they] expect of a Level 59 software engineer role." (DeCaprio Decl. ¶ 14.) As a result, "there is a lot of discretion as to how [managers] reach [their] end results," and the people discussions "var[ied] significantly" based on who was leading the meeting. (Id. ¶¶ 35, 42-43.)

Managers also discussed promotions during people discussions, as they did previously during calibration meetings. They viewed the promotion criteria with a focus on employee impact. (See Helf Decl. ¶ 39.) Managers had "complete discretion" as to their justifications for a promotion. (Id.)

3. Dr. Ryan's Evaluation

To challenge Microsoft's Calibration Process, Plaintiffs submit the expert report of Dr. Ann Marie Ryan, a professor of organizational psychology. (See Ryan Rep. (Dkt. ## 231 (sealed), 379 (redacted)) ¶ 1.) Dr. Ryan reviewed the process Microsoft implemented to make pay and promotions decisions and examined how Microsoft's processes aligned with best practices for employee evaluation. (See id. ¶¶ 5, 11.) After reviewing the relevant materials, Dr. Ryan found "no evidence that compensation and promotion decisions are made reliably at Microsoft." (Id. ¶ 12.) Specifically, Dr. Ryan takes issue with the lack of "sufficient standardization" in the Calibration Process, which "undermines the reliability and validity" of personnel decisions. (Id. ¶ 13.)

Overall, Dr. Ryan criticizes Microsoft for not providing managers sufficient guidance on how to make pay and promotion decisions. For instance, Dr. Ryan observes that Microsoft "does not provide evidence to support how compensation level and promotion decisions are determined," as reflected in the fact that managers are allowed to give open-ended promotion justifications that do not need to be tied to the promotion criteria. (Id. ¶¶ 18-19 (emphasis removed).) As a result, managers may "use varied criteria across individuals for the same decision"; use "irrelevant criteria"; or "use different standards or weighting of the same criteria." (Id. ¶ 20.) Additionally, Dr. Ryan takes issue with the fact that Microsoft did not prescribe any weights to the underlying "what," "how," and "capacity" criteria, so that managers could weigh these criteria differently. (Id. ¶¶ 26-31.)

Moreover, Dr. Ryan observes a lack of uniformity in what information managers considered and the process for considering that information. (See id. ¶ 32.) For example, while Microsoft gave "general guidance" to consider "work products and conversations," "the quantity and nature of information serving as input into compensation level and promotion decisions varied." (Id. ¶ 34.) Dr. Ryan also finds variation in "how the process of evaluating employee performance . . . was conducted" and "how it was determined which employees were discussed." (Id. ¶ 35.)

Lastly, Dr. Ryan critiques Microsoft for inadequately monitoring the Calibration Process to ensure the validity of outcomes. (See id. ¶¶ 40-42.) No reports are generated with regard to monitoring, so it was "unclear" whether the monitoring was consistently done, and if so, the quality of the monitoring efforts. (Id. ¶ 42.) Microsoft allegedly does not train managers on how to weigh criteria in relation to job requirements or how to justify pay and promotion decisions with reference to the criteria used. (Id. ¶ 43.) Nor does Microsoft provide feedback to its managers as to how their decisions are or are not calibrated to the relevant criteria. (Id.)

C. Microsoft Experience with Gender Equity Issues

Microsoft is no stranger to the issue of gender pay equity. In October 2014, the Chief Executive Officer ("CEO") of Microsoft, Satya Nadella, sent a company-wide email stating that "the overall differences in base pay among genders . . . (when we consider level and job title) is consistently within 0.5% at Microsoft." (MCC Docs. at 38.) Microsoft also published an Equal Pay Study in April 2016, in which it stated that women at Microsoft earn 99.80 for every $1.00 earned by men with the same job title and level. See Katherine Hogan, Ensuring Equal Pay for Equal Work, Official Microsoft Blog (Apr. 11, 2016), https://blogs.microsoft.com/blog/2016/04/11/ensuring-equal-pay-equal-work/ (last visited June 21, 2018). After both announcements, Microsoft employees responded that these studies, which only analyzed pay within the same levels, portray a false equity and hide disparity in promotions. (See, e.g., MCC Docs. at 516-17.) As one employee put plainly, "[t]here is an important distinction between equal pay for equal level, and equal pay for equal work." (Id. at 771.)

In response to employee concerns and as acknowledgement of the work remaining to address the pay gap, the Senior Leadership Team ("SLT") at Microsoft "meet[s] on a regular basis" to "review and act on the [D&I] data." (Id. at 1071.) For example, the 2015 SLT D&I Core Priorities Action Plan details several "accountabilities" that various SLT members "own," including measuring and tracking promotion trends and time in role for women; reviewing the recruitment, hiring, and development processes for unconscious bias; updating and requiring all employees to complete an unconscious bias training; and creating sponsorship opportunities for higher-leveled female employees. (Id. at 270-73.) Microsoft also employs ERIT as a tool against discrimination. From 2010 to 2016, ERIT received 238 complaints by female employees, 118 of which concerned gender discrimination. (See Shaver Decl. ¶ 7, Ex. C.) Of these 118, ERIT concluded only once that the complaint was founded—that is, the conduct complained of violated the Anti-Harassment/Anti-Discrimination policy. (See id. at 5.) However, even when ERIT determines a complaint is unfounded, ERIT may still discipline the alleged offender. (See, e.g., Resp. Docs. at 283.)

Additionally, as a federal government contractor, Microsoft has been audited by the Department of Labor's Office of Federal Contract Compliance Programs ("OFCCP") for compliance with federal law and regulations. The OFCCP conducted several audits of Microsoft's locations without findings of violation. (See Parris Decl. ¶ 21, Ex. 19.) On one occasion, however, in 2016, the OFCCP issued a Notice of Violation letter regarding [Redacted] (See id.)

D. Plaintiffs' Employment at Microsoft

Ms. Moussouris and Ms. Muenchow both worked at Microsoft—Ms. Moussouris in the Engineering Profession and Ms. Muenchow in both the Engineering and I/T Operations Professions. Plaintiffs also offer several female employees' declaration testimony detailing their experiences at Microsoft. The court summarizes each in turn.

1. Katherine Moussouris

Ms. Moussouris worked at Microsoft from April 2007 to May 2014. (SAC ¶ 6; Moussouris Decl. (Dkt. # 242) ¶ 2.) Microsoft hired Ms. Moussouris in April of 2007 as a Level 62 Security Program Manager for the Trustworthy Computing Group. (Moussouris Decl. ¶ 2; SAC ¶ 62.) Her position, involving work as a security strategist, featured what Ms. Moussouris calls "unique elements" because she created "new programs that didn't exist before." (Moussouris Dep. (Dkt. # 288-5) at 279:4-11.) Thus, she was not aware of another employee whose role or responsibilities were similar to hers. (See id. at 279:4-17.)

Microsoft promoted Ms. Moussouris to a Level 63 Program Manager II in 2008 and to a Level 64 Senior Program Manager in 2010. (Moussouris Decl. ¶ 2.) After 2010, however, Ms. Moussouris received no more promotions, despite purportedly being "responsible for groundbreaking efforts in the security industry" and obtaining "substantial defensive security research prizes." (See id. ¶¶ 4-5.) In 2012, despite her eligibility for promotion to a Level 65 Principal Program Manager, "two less qualified men in [her] group were promoted over [her], even though they had not performed the same scope of work nor received the same level of recognition in the industry." (Id. ¶ 4.) Her direct manager gave her an initial performance rating of 2 out of 5; her performance rating was bumped down to a 3 during the Calibration Process. (Moussouris Dep. at 68:12-13; 93:10-94:4.) After her performance rating was downgraded, her manager reportedly expressed regret that she "was not given the score that he believes [she] deserved." (Id. at 94:2-4.) Ms. Moussouris believes that gender discrimination affected her 2012 performance review. (Id. at 260:4-6.) The following year, in 2013, Ms. Moussouris was up for a promotion but again passed over; another male colleague secured leadership of the team. (Moussouris Decl. ¶ 4.) Her manager made an initial performance rating recommendation of a 1 out of 5, which was bumped down to a 2 during the Calibration Process. (SAC ¶ 64.)

Ms. Moussouris, as a manager herself, supervised several employees. (See Moussouris Dep. at 147:12-24.) Some of those employees disagreed with her management style. (Id. at 149:10-14, 150:23-25.) As a manager, Ms. Moussouris participated in the Calibration Process by reviewing her employees and attending calibration meetings. (See id. at 103:16-23, 104:19-21.) Some employees under Ms. Moussouris had their initial performance recommendations downgraded. (See id. at 119:4-18.)

Ms. Moussouris describes the Microsoft culture as "hostile towards women." (Moussouris Decl. ¶ 6.) She claims that female employees were "frequently interrupted"; "talked over at meetings"; ignored when they shared ideas that male counterparts would later raise and receive praise for; and excluded from important discussions and business opportunities. (Id.) She also contends that women's judgment and expertise were "much more likely to be called into question than men's." (Id.) Ms. Moussouris describes this problem as "pervasive" and one that female employees and managers regularly discuss. (Id.)

Making complaints to Human Resources ("HR"), Ms. Moussouris attests, "does not make any difference." (Id. ¶ 7.) In fact, Ms. Moussouris made three complaints to HR while at Microsoft. First, she complained about the Director of her group who allegedly sexually harassed women he oversaw. (Id.) Although HR found that he "had in fact sexually harassed these women," the Director was "merely re-assigned to a different group, while retaining his title and influence." (Id.) Before he was re-assigned, he gave Ms. Moussouris a low bonus. (Id.) Ms. Moussouris complained about this retaliation but again, HR took no action. (Id.) Lastly, Ms. Moussouris complained when her new supervisor—a man who was promoted over her—replaced her responsibilities with "low-level tasks that no men in [the] group were asked to do." (Id.) Once again, HR allegedly took no action. (Id.)

2. Holly Muenchow

Ms. Muenchow has worked for Microsoft since September 2002. (Muenchow Decl. (Dkt. # 243) ¶¶ 1-2.) Microsoft initially hired her as a Level 58 Software Test Engineer in the Engineering Profession. (Id. ¶ 2; SAC ¶ 73.) She then became a Level 59 Software Design Engineer in 2005. (Id.) Microsoft promoted Ms. Muenchow to a Level 60 employee in the beginning of 2008 and a Level 61 employee later that same year. (Id.) In 2012, Ms. Muenchow transferred to the I/T Operations Profession to become a Level 62 Senior Program Manager. (Id.) In 2016, she switched back to the Engineering Profession as a Senior Operations Program Manager. (Id.) Ms. Muenchow is currently a Level 63 employee. (See id. ¶ 4.)

Like Ms. Moussouris, Ms. Muenchow believes that Microsoft compensated her less than her similarly-situated male colleagues. (Id. ¶ 3.) Additionally, Ms. Muenchow asserts that Microsoft denied her promotions that were instead given to less qualified men and that the promotions she did receive should have occurred sooner. (Id. ¶ 4; see Muenchow Dep. (Dkt. ## 288-6 (sealed), 287-4 (redacted)) at 60:1-7.) She believes her managers discriminated against her by requiring her to adhere to expectations rooted in gender stereotypes—such as tone of voice—that would not apply to her male colleagues. (Muenchow Dep. at 52:17-53:6, 55:1-12.) This gender bias in expected behavior, she asserts, informed much of the negative feedback she received and resulted in her achievements not being acknowledged as her male colleagues' efforts were. (See, e.g., id. at 177:10-16, 190:25-191:10.) Although Ms. Muenchow never advanced higher than a Level 63, she saw "male colleagues be promoted to high levels for which [she] was qualified and not considered." (Muenchow Decl. ¶ 4.) She also observed "men generally advanc[ing] more rapidly than women, despite similar scope of work and performance." (Id. ¶ 5.)

Ms. Muenchow also describes a culture of hostility at Microsoft towards women. (See id. ¶ 6.) She states that women are "held to a different standard than men." (Id.) For instance, women are labeled as "too aggressive" when they speak up in meetings, whereas men "routinely interrupt or talk over women without criticism." (Id.) Her former manager, who was female, revealed to Ms. Muenchow that she could not "advocat[e] passionately" for her employees during calibration meetings because of this perceived bias. (Id.)

Like Ms. Moussouris, Ms. Muenchow has also filed complaints with HR to no avail. (See id. ¶ 7.) Ms. Muenchow complained to HR about the "culture of bias towards women." (Id.) In response, HR informed her that Microsoft would introduce "bias training," but Ms. Muenchow reports that the training has not resolved the issue. (Id.) Ms. Muenchow has also requested to see Microsoft's data regarding compensation and promotion rates for female employees, but HR stated that it did not have access to that data. (Id.)

3. Other Employees

Nine other female employees echo Ms. Moussouris and Ms. Muenchow's sentiments. (See, e.g., Dove Decl. (Dkt. # 239) ¶ 4.) The female employees, through declarations submitted by Plaintiffs, report similar claims of compensation discrimination, contending that they were paid less than their similarly-situated male counterparts (see Miller Decl. (Dkt. # 241) ¶ 5) and at times, even below the pay of less-qualified male colleagues (see Smith Decl. (Dkt. # 244) ¶ 5). At least one woman, Debra Dove, reports that her pay disparity resulted from the Calibration Process, during which "someone had to receive zero rewards, and that someone became [her] because of [her] gender." (Dove Decl. ¶ 5.)

These female employees also report that they were denied promotions ultimately given to similarly-situated or less qualified male employees. (See, e.g., Albert Decl. (Dkt. # 237) ¶ 6; Smith Decl. ¶ 6.) For instance, Heidi Boeh sought a promotion after returning from maternity leave. (Boeh Decl. (Dkt. # 238) ¶ 5.) Her manager stated that he "did not want to 'waste' a promotion on [her] in case [she] became pregnant again," due to his understanding that she "would want to 'time' [her] children to be close in age." (Id.) Olga Hutson inquired about a promotion but "was told repeatedly that it was not possible . . . to receive a promotion to Level 65 because almost no one made it to that level." (Hutson Decl. (Dkt. # 240) ¶ 5.) But, while she remained at a Level 64 for her six-year tenure, Microsoft promoted several men to Level 65. (Id.) Suzanne Sowinska, while participating as a manager in the Calibration Process, noticed that women were "disadvantaged compared to men with no greater qualifications." (Sowinska Decl. (Dkt. # 245) ¶ 5.) Male employees' projects were "valued more highly than similar projects managed by women . . . even when the technical difficulty and value to the company was greater for the women's projects." (Id.)

Moreover, the nine declarants describe Microsoft's culture of hostility towards women, leading female employees to feel undervalued and marginalized. Many describe observations of sexual harassment, including inappropriate comments regarding female employees' looks, figures, or clothing; groping and other unwanted touching; and other characteristics of a "male-dominated culture" within which female employees struggled to operate. (Sowinska Decl. ¶ 7; see also, e.g., Warren Decl. (Dkt. # 248) ¶ 7 (describing a Microsoft party with "scantily clad women dancing on tables"); Miller Decl. ¶ 8 (describing inappropriate comments regarding her clothing, unwanted touching, and manager inquiries into her marriage status).) Some were harassed themselves. (See, e.g., Boeh Decl. ¶ 7.) Mary Smith described how a male colleague "screamed and cursed . . . and threatened to kill [her]." (Smith Decl. ¶ 8.) When she informed her manager, he acknowledged that the male colleague was "sexist" but did nothing further. (Id.)

Almost all of the female employees recount how Microsoft excluded them from business opportunities and applied double-standards—rooted in gender stereotypes about behavior—to dictate how female employees should behave. (See, e.g., Vaughn Decl. (Dkt. # 247) ¶ 6.) For instance, Jennifer Underwood states Microsoft did not give her the staff, support, and funding to attend industry events while male peers received "significant financial and administrative support to fulfill the same job duties." (Underwood Decl. (Dkt. # 246) ¶ 7.) Microsoft allegedly downgraded Ms. Sowinska's performance rating in the Calibration Process because she "did not smile enough." (Sowinska Decl. ¶ 7.) Kristen Warren observed that "men are praised for exhibiting strong opinions and being assertive, while women are admonished for the same behavior." (Warren Decl. ¶ 7.) In fact, managers told her to "control [her] 'emotions' after [she] expressed opinions in meetings, even though men making similar comments did not receive negative feedback." (Id.)

None of the female employees believes that the HR complaint process is effective. Many complained to HR regarding their lack of promotions (Vaughn Decl. ¶ 7), their performance ratings (Dove Decl. ¶ 9), or their inability to secure necessary resources (Underwood Decl. ¶ 9). HR allegedly responded that there was nothing it could do. (See, e.g., Vaughn Decl. ¶ 7.) Other women complained of their treatment upon returning from maternity leave, but HR failed to respond or follow up. (See Albert Decl. ¶ 8; Boeh Decl. ¶ 8.) HR similarly did not investigate complaints regarding threats or harassment. (See Smith Decl. ¶ 9.) When HR did follow up, female employees were dissatisfied with the ensuing investigation. For instance, when investigating Ms. Hutson's complaint, HR acknowledged that the manager "acted inappropriately" but nonetheless found no violation of Microsoft's Anti-Harassment/Anti-Discrimination policy. (Hutson Decl. ¶ 6.) The female employees also state that female complainants face retaliation. Ms. Smith recounts that after complaining about threats and hostility from male colleagues, Microsoft assigned her responsibilities outside the normal scope of her position. (Smith Decl. ¶ 10.)

E. Procedural History

Plaintiffs filed their original complaint on September 16, 2015, challenging Microsoft's continuing policy, pattern, and practice of sex discrimination against female employees in technical and engineering roles. (Compl. (Dkt. # 1) ¶ 1.) Plaintiffs amended their complaint on October 27, 2015 (see FAC (Dkt. # 8)), and Microsoft filed a motion to dismiss for failure to state a claim, or alternatively, a motion for a more definite statement (MTD (Dkt. # 23) at 1). The court granted in part and denied in part Microsoft's motion. (See MTD Order (Dkt. # 52) at 2.) The court denied the motion as to Plaintiffs' disparate treatment claims and Ms. Moussouris's retaliation claims (id. at 20, 24-28) but dismissed, with leave to amend, Plaintiffs' disparate impact claims for failing to allege "any facts from which the court can plausibly infer the allegedly offending employment practice causes the alleged disparate impact" (id. at 21, 29).

Plaintiffs amended the complaint on April 6, 2016. (See SAC.) This second amended complaint described in detail the Calibration Process (id. ¶¶ 27-32, 37-40) and asserted that Microsoft systematically undervalued female technical employees in this process because female employees received, on average, lower rankings despite equal or better performance (id. ¶ 34). Microsoft again moved to dismiss Plaintiffs' disparate impact claims. (See 2d MTD (Dkt. # 62).) This time, the court denied the motion, noting that Plaintiffs sufficiently identified a challenged employment practice (2d MTD Order (Dkt. # 134) at 9) and alleged sufficient facts for the court to draw the reasonable inference that Microsoft's Calibration Process led to arbitrary differentiations between employees and disparately impacted women (id. at 10).

The Plaintiffs subsequently filed their motion for class certification. (See MCC.) Both parties then submitted motions to exclude various experts proffered in support of and in opposition to class certification. (See Farber MTE (Dkt. # 362); Saad MTE (Dkt. # 364); Young MTE (Dkt. ## 367 (sealed), 368 (redacted)).) The court denied Microsoft's motion to exclude the expert opinion of Dr. Henry S. Farber and the majority of Plaintiffs' motion to exclude the expert opinion of Dr. Ali Saad. (MTE Order (Dkt. # 467) at 16-23, 29-33.) However, the court excluded (1) Dr. Saad's "business need" word search analysis because it was not the product of reliable principles and methods (id. at 26-28); and (2) Rhoma Young's evaluation of the ERIT process because it was not based on sufficient facts or data (id. at 33-39). The court now considers the class certification motion.

III. ANALYSIS

Plaintiffs move to certify a nationwide class of female Microsoft employees, leveled 59 to 67, who worked in the Engineering and/or I/T Operations Professions from September 16, 2012, to the present. (MCC at 1.) All class members have a claim of discrimination in pay, and class members in Levels 60 to 64 have claims for discrimination in promotion. (Id.) Plaintiffs seek certification under Federal Rule of Civil Procedure 23(b)(2) for their injunctive relief claims, Rule 23(b)(3) for their damages claims, and Rules 23(b)(2), 23(b)(3), and/or 23(c)(4) for certification of the liability phase. (Id. at 2-3); see Fed. R. Civ. P. 23(b)-(c).

Microsoft not only opposes class certification on the merits but also seeks to strike portions of Plaintiffs' reply brief and accompanying documents. (See Resp.; Surreply.) The court addresses the surreply before moving to the class certification issue.

A. Microsoft's Surreply

Pursuant to Local Civil Rule 7(g), Microsoft moves to strike the following materials connected with Plaintiffs' reply: (1) legal argument in the declaration of Plaintiffs' counsel; (2) Dr. Ryan's reply report; and (3) a footnote in the reply that makes allegedly unsupported allegations regarding pay discrepancies. (Surreply at 1.) The court addresses each in turn.

First, Microsoft challenges as improper legal argument Plaintiffs' chart entitled "Microsoft's alleged variation in the pay and promotion process is not material and/or is not supported by the evidence it cites to." (See Reply Shaver Decl. (Dkt. ## 343 (sealed), 359-2 (redacted)) ¶ 4, Ex. A ("Reply Chart") at 1.) In Plaintiffs' reply, they argue that "Microsoft offers a few immaterial variations in how pay or promotion decisions were internally documented or communicated across groups." (Reply (Dkt. ## 342 (sealed), 359-1 (redacted)) at 3.) After addressing one example, Plaintiffs point the court to their chart that addresses how the remaining variations are "immaterial and/or unsupported by the record." (Id.) For instance, in regards to Microsoft's assertion that each team handled calibration meetings differently, Microsoft relies in part on deposition testimony regarding how HR members varied in their documentation of calibration meetings. (Reply Chart at 7.) Plaintiffs state in their chart that the "presence or absence of other HR team members in the room, or the identity of the person documenting common information, are irrelevant to the common criteria challenged in this action." (Id.)

The court agrees that Plaintiffs' chart should be stricken as improper legal argument outside the court-approved page limit. (See id. at 1-2.) "Declarations . . . should not be used to make an end-run around the page limitations . . . by including legal arguments outside of the briefs." King Cty. v. Rasmussen, 299 F.3d 1077, 1082 (9th Cir. 2002) (citing Fed. R. Civ. P. 56(e)). Plaintiffs' chart, while couched in factual assertions, contains significant argument regarding why several of Microsoft's assertions are erroneous. (See, e.g., Reply Chart at 7.) Such argument goes beyond what is appropriate for a declaration. See Fed. R. Civ. P. 56(e); see also Sierra Club v. BNSF Ry. Co., No. C13-0976JCC, 2017 WL 3141899, at *1 (W.D. Wash. July 25, 2017) (striking legal arguments in declarations that were "an attempt to circumvent the reply brief page limits"). Thus, the court will not consider Plaintiffs' chart in its determination of the class certification motion.⁴

4 Microsoft also seeks to strike paragraph 5 of Plaintiffs' counsel's declaration, in which she attests to the law firm Lane Powell's involvement as counsel for both class member declarants and Microsoft. (See Surreply at 2 (citing Reply Shaver Decl. ¶ 5).) This paragraph sets out factual allegations regarding Lane Powell's representation of putative class members during their depositions. (See Shaver Decl. ¶ 5, Ex. B (copies of Lane Powell engagement letters).) These factual allegations are appropriate for a declaration. Indeed, Microsoft does not identify how this paragraph constitutes legal argument in its one-sentence argument on this point. (See Surreply at 2.) Thus, the court declines to strike the paragraph.

Second, Microsoft seeks to strike Dr. Ryan's reply report as an improper rebuttal report. The court agrees. Plaintiffs offer Dr. Ryan's "Reply Expert Report" to "respond to the portions of Microsoft's [response] . . . that address the topics of [her original report]." (Ryan Reply (Dkt. # 359-11) ¶ 1.) But rebuttal reports are proper only "'to contradict or rebut evidence on the same subject matter' from an opposing party's expert." Theoharis v. Rongen, No. C13-1345RAJ, 2014 WL 3563386, at *1 (W.D. Wash. July 18, 2014) (quoting Fed. R. Civ. P. 26(a)(2)(D)(ii)). Plaintiffs readily point out that Microsoft "leaves Dr. Ryan's opinions unrebutted, offering no expert opinion opposing hers." (Reply at 4.) Thus, Dr. Ryan's additional report "attempts to augment // // [Plaintiffs'] reply by attacking a legal argument from [Microsoft's] counsel."⁵ See Johnson v. Hartford Cas. Ins. Co., No. 15-cv-04138-WHO, 2017 WL 2224828, at *3 (N.D. Cal. May 22, 2017). As Plaintiffs' counsel conceded in oral argument, such use of a rebuttal report is improper.⁶ Accordingly, the court strikes Dr. Ryan's reply report and does not consider it further.

5 Even though Dr. Ryan's reply report mentions Ms. Young's evaluation of ERIT, the reply report centers on rebutting Microsoft's arguments—not on Ms. Young's expert report. For example, Dr. Ryan addresses how Microsoft witnesses support her conclusions (Ryan Reply Rep. ¶ 3) and how Microsoft's assertions regarding training are inaccurate (id. ¶ 4). Indeed, Dr. Ryan's reply report only briefly mentions Ms. Young's report. (See id. ¶ 4.b.) Thus, Dr. Ryan's reply report is not a rebuttal of Ms. Young's report.

6 This is true even though the report is labeled as a "reply report." See K&N Eng'g, Inc. v. Spectre Performance, No. EDCV 09-1900CAP (DTBx), 2011 WL 13131157, at *9 (C.D. Cal. May 12, 2011) (excluding an expert's reply report because neither the scheduling order nor the Federal Rules of Civil Procedure "permit 'reply' reports").

Third, and lastly, Microsoft seeks to strike a footnote in Plaintiffs' reply that makes pay discrepancy allegations. (Surreply at 3.) Specifically, Plaintiffs state that "[c]lass member[] witnesses have been paid less than comparable male coworkers" but provide no citation for that statement. (See Reply at 17 n.35.) Microsoft contends that the court should strike this unsupported allegation. The court disagrees. The court may disregard arguments of counsel in briefing if they are "wholly unsupported" or if they "conflict[] with the available evidence." Wilbur v. City of Mount Vernon, No. C11-1100RSL, 2012 WL 600727, at *2 (W.D. Wash. Feb. 23, 2012). However, arguments can also be based on "common sense inferences or limited/equivocal // // evidence." Id. Thus, Plaintiffs' lack of citation for their statement may affect the weight afforded, but it is unnecessary to strike the footnote.⁷ See id.

7 Moreover, contrary to Microsoft's contentions (see Surreply at 3), evidence does not need to be admissible to be considered at the class certification stage, see Sali v. Corona Reg'l Med. Ctr., --- F.3d ----, 2018 WL 2049680, at *5 (9th Cir. May 3, 2018). Thus, unlike the summary judgment cases Microsoft cites (see Surreply at 3), the court is not limited to considering only admissible evidence.

Accordingly, the court strikes the chart attached to Plaintiffs' counsel's declaration and Dr. Ryan's reply report. The remainder of the reply stands as filed.

B. Class Certification Motion

Having determined the materials properly before the court, the court now turns to the Plaintiffs' motion for class certification. "The class action is 'an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only.'" Comcast Corp. v. Behrend, 569 U.S. 27, 33 (2013) (quoting Califano v. Yamasaki, 442 U.S. 682, 700-01 (1979)). "Class certification is governed by Federal Rule of Civil Procedure 23." Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 345 (2011). Under Rule 23(a), the party seeking certification must first demonstrate that:

(1) the class is so numerous that joinder of all members is impracticable;
(2) there are questions of law or fact common to the class;
(3) the claims or defenses of the representative parties are typical of the claims or defenses of the class; and
(4) the representative parties will fairly and adequately protect the interests of the class.

Fed. R. Civ. P. 23(a). Rule 23(a)'s four subparts are generally referred to as the requirements of numerosity, commonality, typicality, and adequacy of representation. Certification is proper "only if 'the trial court is satisfied, after a rigorous analysis, that the prerequisites of Rule 23(a) have been satisfied.'" Dukes, 564 U.S. at 350-51 (quoting Gen. Tel. Co. of Sw. v. Falcon, 457 U.S. 147, 161 (1982)).

In addition to meeting the Rule 23(a) prerequisites, the action must also fall into one of three categories under Rule 23(b). Id. at 349. For their injunctive relief claims, Plaintiffs rely on Rule 23(b)(2), which applies when "the party opposing the class has acted or refused to act on grounds that apply generally to the class, so that the final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole." Fed. R. Civ. P 23(b)(2); (see MCC at 36-38.) "[T]he key to the (b)(2) class is the indivisible nature of the injunctive or declaratory remedy warranted." Ellis v. Costco Wholesale Corp., 657 F.3d 970, 987 (9th Cir. 2011) ("Ellis I") (internal quotation marks omitted) (quoting Dukes, 564 U.S. at 360). Plaintiffs additionally rely on Rule 23(b)(3) for their damages claim. (MCC at 38-43.) Rule 23(b)(3) applies when a "question of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy." Fed. R. Civ. P. 23(b)(3).

"[S]ometimes it may be necessary for the court to probe behind the pleadings before coming to rest on the certification question." Dukes, 564 U.S. at 350 (internal quotation marks omitted) (quoting Falcon, 457 U.S. at 160). "The class determination generally involves considerations that are enmeshed in the factual and legal issues comprising the plaintiff's cause of action." Id. at 351 (internal quotation marks omitted) (quoting Falcon, 457 U.S. at 160). Nevertheless, at this stage, the court "is merely to decide a suitable method of adjudicating the case" and "should not turn class certification into a mini-trial on the merits." Edwards v. First Am. Corp., 798 F.3d 1172, 1178 (9th Cir. 2015) (internal quotation marks omitted) (quoting Ellis I, 657 F.3d at 983 n.8).

"The party seeking class certification bears the burden of establishing that the proposed class meets the requirements of Rule 23." Id. at 1177. "A party seeking class certification must affirmatively demonstrate his compliance with [Rule 23]—that is, he must be prepared to prove that there are in fact sufficiently numerous parties, common questions of law or fact, etc." Dukes, 564 U.S. at 351; Behrend, 569 U.S. at 33 (requiring the party seeking certification to affirmatively demonstrate compliance with Rule 23 "through evidentiary proof"). The court may grant certification only after "a rigorous analysis . . . [determining] that the prerequisites of Rule 23(a) have been satisfied." Dukes, 564 U.S. at 350-51. If a court is not fully satisfied that the requirements of Rules 23(a) and (b) have been met, it should deny certification. Falcon, 457 U.S. at 161.

Of the four Rule 23(a) prerequisites, Microsoft challenges commonality, typicality, and adequacy.⁸ (See Resp. at 15-41.) The court concludes that Plaintiffs have not affirmatively demonstrated that common questions of law or fact exist across the proposed class; that Plaintiffs' claims are typical of the absent class members' claims; or that Plaintiffs will fairly and adequately protect the interests of the class. The court addresses each prerequisite in turn. //

8 Microsoft does not challenge numerosity, and the court finds that the Plaintiffs' proposed class, consisting of over 8,000 putative members, satisfies the numerosity requirement. See Dunakin v. Quigley, 99 F. Supp. 3d 1297, 1326-27 (W.D. Wash. 2015) (accepting that 40 or more class members will generally satisfy numerosity).

1. Commonality

As is true in many Title VII class actions, the "crux of this case is commonality." See Dukes, 564 U.S. at 349. Rule 23(a)(2) requires a plaintiff to show that "there are questions of law or fact common to the class." Fed. R. Civ. P. 23(a)(2). The requirements of commonality have been "construed permissively"; not all questions of fact and law need to be common. Ellis I, 657 F.3d at 981. However, because any competently crafted class complaint "literally raises common 'questions,'" not every common question suffices. Dukes, 564 U.S. at 349; Ellis I, 657 F.3d at 981. For instance, reciting questions such as "Do our managers have discretion over pay" or "Is that an unlawful employment practice" are insufficient to obtain class certification. Dukes, 564 U.S. at 349. Instead, Plaintiffs' claims must "depend upon a common contention" that is "of such a nature that it is capable of classwide resolution—which means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke." Id. at 350. Put another way, Plaintiffs must pose a "common question" that "will connect many individual promotional decisions to their claim for class relief" and "produce a common answer to the crucial question why was I disfavored?" Ellis I, 657 F.3d at 981 (internal quotation marks omitted) (quoting Dukes, 564 U.S. at 352).

At the outset, the court notes that from the time of their briefing to oral argument, Plaintiffs' theory and arguments surrounding commonality have shifted. In their briefing, Plaintiffs make no distinction between their pay and promotion claims, arguing that Microsoft's "common but unvalidated criteria," "set of common procedures," and the "final approval by just four EVPs" cabined the discretion that lower managers exercised in both pay and promotion determinations. (See Reply at 12-13; see also MCC at 32-34 (making no distinction between pay and promotion in their argument on commonality).) During oral argument, however, Plaintiffs presented a new theory that suggested different analyses for pay and promotion: They maintained that the use of Stock Levels removed discretion for pay determinations whereas the common criteria and procedures cabined discretion for promotion determinations.⁹ "It is inappropriate to present a new argument at oral argument and deny the [c]ourt and opposing counsel a chance to review the merits of such an argument." Value Home Auctions, Inc. v. X-Wire Techs., Inc., No. SACV 10-0153 AG (RNBx), 2011 WL 13225021, at *4 (C.D. Cal. Mar. 23, 2011). Nonetheless, the court considers all of Plaintiffs' arguments and addresses Plaintiffs' disparate impact and disparate treatment claims in turn.¹⁰ // //

9 Tellingly, when asked where in their briefing Plaintiffs made this argument regarding Stock Level, Plaintiffs pointed only to the briefing's background section. (See MCC at 8-9 (justifying Dr. Farber's failure to control for Stock Level in his statistical analysis); id. at 12 [Redacted] id. at 13-16 (citing employee complaints about Stock Level).) Nowhere in the cited pages do Plaintiffs present Stock Level as a common corporate policy that eliminates the discretion surrounding pay determinations.

10 The court recognizes that the distinction between a disparate impact claim and a disparate treatment claim, for commonality purposes, is often illusory because the "actual argument in support of class certification ultimately makes little distinction between the two." Dukes v. Wal-Mart Stores, Inc., 964 F. Supp. 2d 1115, 1119 (N.D. Cal. 2013) ("Dukes II"); see also Jones v. Nat'l Council of Young Men's Christian Ass'ns, 34 F. Supp. 3d 896, 909 (N.D. Ill. 2014) (recognizing overlap in the analysis of disparate impact and disparate treatment claims because of the "common cause for the injuries claimed"). Nonetheless, the court addresses each claim separately to fully flesh out its rationale, with the understanding that the analysis for the two intersects to some degree.

a. Disparate Impact Commonality

Disparate impact claims under Title VII challenge "a facially neutral policy or practice that causes a disparate impact on a protected group, even if the employer has no intent to discriminate." Williams v. Boeing Co., 225 F.R.D. 626, 634 (W.D. Wash. 2005). To establish a prima facie case of disparate impact, plaintiffs must: (1) show a significant disparate impact on a protected class or group; (2) identify the specific employment practices at issue; and (3) show a causal relationship between the challenged practices and the disparate impact. Hemmings v. Tidyman's Inc., 285 F.3d 1174, 1190 (9th Cir. 2002). If plaintiffs establish a prima facie case, the burden shifts to the employer to show that its challenged practices are consistent with business necessity and that there was no less discriminatory alternative. Ellis v. Costco Wholesale Corp., 285 F.R.D. 492, 505 (N.D. Cal. 2012) ("Ellis II").

The court first reviews the comprehensive legal landscape governing commonality for disparate impact claims implicating managerial discretion. The court then applies that law to the circumstances here.

i. Law Governing Disparate Impact Commonality

Wal-Mart Stores, Inc. v. Dukes directly addressed the issue of commonality in a Title VII gender discrimination case that challenged the discretion exercised by individual supervisors over pay and promotion matters. See 564 U.S. at 344-45. The Dukes plaintiffs alleged that "local managers' discretion over pay and promotions [was] exercised disproportionately in favor of men, leading to an unlawful disparate impact on female employees." Id. at 344-45. Plaintiffs' theory of commonality was that Wal-Mart's "strong and uniform 'corporate culture' permits bias against women to infect . . . the discretionary decisionmaking of each one of Wal-Mart's thousands of managers—thereby making every woman at the company the victim of one common discriminatory practice." Id. at 345.

The Supreme Court observed that "[t]he only corporate policy that the plaintiffs' evidence convincingly establishes is Wal-Mart's 'policy' of allowing discretion by local supervisors over employment matters." Id. at 355. Such a policy of discretion is, "[o]n its face . . . just the opposite of a uniform employment practice that would provide the commonality needed for a class action; it is a policy against having uniform employment practices." Id. Although the Court recognized that in appropriate cases, granting unlimited discretion to lower-level supervisors could be the basis of Title VII liability, that possibility of liability "does not lead to the conclusion that every employee in a company using a system of discretion has such a claim in common." Id. In fact, the Court surmised, the exact opposite is likely true:

[L]eft to their own devices most managers in any corporation—and surely most managers in a corporation that forbids sex discrimination—would select sex-neutral, performance-based criteria for hiring and promotion that produce no actionable disparity at all. Others may choose to reward various attributes that produce disparate impact—such as scores on general aptitude tests or educational achievements. And still other managers may be guilty of intentional discrimination that produces a sex-based disparity.

Id. Because "demonstrating the invalidity of one manager's discretion will do nothing to demonstrate the invalidity of another's," a party basing commonality on such a system of discretion "will be unable to show that all the employees' Title VII claims will in fact depend on the answers to common questions." Id. at 355-56.

But Dukes did not entirely foreclose the ability to establish commonality when the employer operates under a policy allowing discretion. See Ellis II, 285 F.R.D. at 518 (noting that some discretion "does not in and of itself preclude class certification"); accord Scott v. Family Dollar Stores, Inc., 733 F.3d 105, 113-14 (4th Cir. 2013) (describing how a disparate impact claim challenging discretion may satisfy commonality post-Dukes). Plaintiffs challenging a system of discretion must identify "a common mode of exercising discretion that pervades the entire company." Dukes, 564 U.S. at 356. In other words, when plaintiffs wish to challenge numerous employment decisions at once, they must point to "some glue holding the alleged reasons for all those decisions together." Id. at 352. Otherwise, it would be "quite unbelievable that all managers would exercise their discretion in a common way without some common direction." Id. at 356. The Court determined that both the plaintiffs' aggregated statistical evidence and anecdotal evidence fell "well short" of indicating a "common direction" from upper management, and thus, the plaintiffs failed to demonstrate the existence of common issues. Id. at 356-58.

On remand, the Dukes plaintiffs narrowed their class significantly, from a nationwide class to three regional classes, all within California. Dukes II, 964 F. Supp. 2d at 1118-19. They additionally identified several corporate policies, such as a "tap on the shoulder" system for promotions and company-wide guidelines for pay decisions, which allegedly provided "common direction." Id. at 1119. For instance, plaintiffs alleged that Wal-Mart limited the managers' discretion by requiring them to consider various criteria, such as "communication skills" or "ability to learn." Id. at 1126. The district court rejected the plaintiffs' characterization. The court concluded that the identified criteria "were so vague or numerous that they imposed no real constraints," and additionally, Wal-Mart did not prohibit managers from considering other factors of their own choosing. Id.

Thus, although plaintiffs identified some company-wide policies, the court noted that the plaintiffs were not challenging the policies themselves or the faithful application of those policies; instead, at its core, they continued to challenge the "broad discretion managers retain in applying the vague criteria." Id. at 1126-27. Put another way, the plaintiffs' argument still boiled down to the theory that "managers, who were left without meaningful guidance in applying the impossibly vague criteria, fell back on their own stereotyped views of women in making pay and promotion decisions." Id. at 1127. While this may be "a perfectly logical theory" for liability, the court observed that it "leaves [p]laintiffs right back where they started: challenging Wal-Mart's practice of delegating discretion to local managers, which the Supreme Court specifically held was not a specific employment practice supplying a common question sufficient to certify a class." Id.

Like Dukes II, other courts addressing similar challenges have concluded that commonality is lacking. See, e.g., Tabor v. Hilti, Inc., 703 F.3d 1206, 1229 (10th Cir. 2013) (recognizing that courts "have generally denied certification when allegedly discriminatory policies are highly discretionary"). In Jones v. National Council of Young Men's Christian Associations, the plaintiffs argued that the company did not have "effective oversight" and thus allowed local managers' exercise of discretion to result in adverse outcomes for the protected class. 34 F. Supp. 3d at 905. Plaintiffs attempted to "downplay[] the degree of discretion" by describing the entire evaluation and promotion process as a mandatory company-wide policy. Id. at 904. The Jones court rejected this attempt to "increas[e] the level of abstraction in defining the policy." Id.; see id. at 905 (observing that at the level of abstraction the plaintiffs propose, "every company—even Wal-Mart—could be said to have a company-wide policy"). Instead, the court characterized the plaintiffs' challenge to oversight as "precisely the argument held to be inadequate to support class certification in [Dukes]." Id.; see also Bolden v. Walsh Constr. Co., 688 F.3d 893, 898 (7th Cir. 2012) ("[Dukes] tells us that local discretion cannot support a company-wide class no matter how cleverly lawyers may try to repackage local variability as uniformity.").

Accordingly, post-Dukes, the crux of the commonality inquiry for a system of discretion lies in whether lower-level supervisors operate under "a common mode of exercising discretion"—put differently, whether some company-wide policy provides sufficient "common direction" such that individual exercises of discretion nonetheless produce a common answer to the question "why was I disfavored." See 564 U.S. at 352, 356 (emphasis removed); see also Ross v. Lockheed Martin Corp., 267 F. Supp. 3d 174, 198 (D.D.C. 2017) (requiring the plaintiffs to demonstrate that "all managers would exercise their discretion in a common way"). Courts considering the issue analyze several factors in determining whether a common mode of exercising discretion pervades the entire company. Those factors include: (1) the nature of the purported class; (2) the // process through which discretion is exercised; (3) the criteria governing the discretion; and (4) the involvement of upper management.

First, the nature of the purported class is, at the very least, relevant. Although class size "has no per se bearing on commonality," courts recognize that "when the claims focus in part on the exercise of managerial discretion, it is reasonable to suspect that the larger the class size, the less plausible it is that a class will be able to demonstrate a common mode of exercising discretion." Ellis II, 285 F.R.D. at 509. Thus, a "more centralized, circumscribed environment generally increases . . . the consistency with which managerial discretion is exercised." Brown v. Nucor Corp., 785 F.3d 895, 910 (4th Cir. 2015); see also Rollins v. Traylor Bros., Inc., No. C14-1414JCC, 2016 WL 258523, at *7 (W.D. Wash. Jan. 21, 2016), abrogated on other grounds, 2016 WL 5942943 (W.D. Wash. May 3, 2016). In Brown, for example, the proposed class consisted of about 100 putative members, all located in a single facility. Id. at 910. For a "localized, circumscribed class" of that size, a policy of subjective, discretionary decision-making can "more easily form the basis of Title VII liability." Id. at 916; see also Chen-Oster v. Goldman, Sachs & Co., --- F.R.D. ----, 2018 WL 1609267, at *2 (S.D.N.Y. Mar. 30, 2018) (approving class of 1,762 to 2,300 people holding two job positions). On the other hand, for a nationwide class, "proving a consistent exercise of discretion will be difficult, if not impossible." Brown, 785 F.3d at 916.

Second, courts look to what procedures govern the discretion and analyze the rigidity of the process through which discretion is exercised. Simply showing a process in which discretion would be exercised is insufficient to establish commonality. See Valerino v. Holder, 283 F.R.D. 302, 313 (E.D. Va. 2012). In Valerino, the company set up an evaluation and promotions process in which employees underwent several phases of evaluation by local supervisors. Id. at 304-07. Plaintiffs argued that this company-wide promotion process created a common mode of exercising discretion. Id. at 313. The court disagreed. Although the court acknowledged that the company "set[] up a structure," that structure alone was insufficient because the process merely laid out "points at which particular people exercise discretion." Id. The court remarked that "[c]ertainly every application is subject to that structure, but so was every applicant in Wal-Mart subject to a [promotions] structure." Id. at 313-14. Nothing about the process constrained the exercise of discretion; thus, "[t]hat discretion itself cannot be said to be the same or 'common.'" Id. at 313.

In contrast, a structure that imposes specific requirements, beyond simply laying out general steps in a process, may show sufficient common direction. The Ellis II court noted that the company not only had a single procedure that controlled the people generally involved and the process by which they made decisions, but the company also imposed several non-negotiable requirements, such as mandatory promotion from within the company and a prohibition on posting job openings. See 285 F.R.D. at 511. Those requirements—implemented across the company—distinguished the discretion at issue in Ellis II from the unfettered discretion in Dukes. See id. at 509; see also McReynolds v. Merrill Lynch, Pierce, Fenner & Smith, Inc., 672 F.3d 482, 490 (7th Cir. 2012) (holding that two company-wide policies that required the managers to allow certain practices, // rather than leaving managers to make the decision themselves, sufficiently restrained discretion).

Third, courts consider the criteria that govern the discretionary decisions. Again, whether a set of criteria creates a common mode of exercising discretion depends on the rigidity of that criteria. Subjective criteria, prone to different interpretations, generally do not provide common direction. See, e.g., Ross, 267 F. Supp. 3d at 198. In Ross, "[t]he mere fact that all . . . supervisors used the same allegedly ill-defined numerical rubric . . . says nothing about how individual supervisors exercised what discretion was left to them." Id. Similarly, in Jones, managers had to consider "amorphous concepts," such as "the organization's strategy, the employee's role in the organization, and the market value of the employee's job," as criteria for determining pay and promotions. 34 F. Supp. 3d at 906. "That supervisors evaluate candidates according to specific, but subjective factors . . . does not make the decisions produced by the process meaningfully less discretionary." Id.

On the other hand, objective criteria or criteria commonly understood throughout the company may establish a common mode of exercising discretion. In Ellis II, the court observed that most of the criteria, such as the applicant's ability to relocate or an applicant's prior experience in the position, was objective and thus not open to interpretation by individual decision-makers. See 285 F.R.D. at 514. The remaining criteria, while subjective, carried definitions within the company that were uniformly understood. See id. For instance, the CEO attested that everyone in the company had a "pretty clear understanding" of the company criteria and denied that this criteria "varied based on the whims of local managers." Id. Thus, because even the subjective criteria was interpreted in a uniform manner, the Ellis II court found that the criteria for promotion differed from that in Dukes where "managers appl[ied] their own subjective criteria when selecting candidates." Id. at 518.

Fourth, and lastly, the involvement of top management in the discretionary decision-making is a key consideration. See Dukes, 564 U.S. at 350 (suggesting that commonality may be satisfied if the class were limited to one supervisor); Brown, 785 F.3d at 916 (recognizing commonality if high-level personnel exercise the discretion at issue). If, for instance, a single decision-maker or a "small, cohesive group" vetted all of the pay and promotion decisions, then the involvement of that one individual, or small group of individuals could constitute a common practice. Jones, 34 F. Supp. 3d at 908. In In re Johnson, "every promotion decision was ultimately made by the Director of [the company at issue]." 760 F.3d 66, 73 (D.C. Cir. 2014). Thus, although different lower-level decision-makers may have injected some subjectivity, the fact that the evaluation process ended with one individual provided the necessary "common contention" capable of class-wide resolution. See id.; see also Chen-Oster, 2018 WL 1609267, at *4-5 (certifying a class where the upper management developed the list of eligible candidates and "[u]ltimately, [the] management committee decide[d] who [wa]s promoted").

Similarly, the Ellis II court focused on the "consistent[] and pervasive" involvement of top management in the promotions process. 285 F.R.D. at 511-12. Senior management chose the pool of candidates eligible for promotions by regularly assessing potential candidates during "floor walks." Id. at 512-14. The CEO conceded that he gave "[a] lot" of instruction on how to fill the positions, id. at 512, including "personally instruct[ing] his staff as to the criteria" for promotions, id. at 514. Because top management was intimately involved in both designing and executing the promotion process, the court determined that even though lower-level managers had input regarding promotion candidates, top management "actually ma[d]e the decisions." Id. at 515.

But mere approval of decisions by higher-level executives, without more, falls short. When "evidence establishes that the recommendations of lower level . . . managers [are] almost always accepted," this "limited oversight" does not establish top management as a common denominator. Jones, 34 F. Supp. 3d at 908. In Jones, plaintiffs pointed to "no evidence showing that any member of senior management changed pay or promotion recommendations . . . with any frequency, much less that they did so with [] regularity." Id. This minimal level of involvement pales in comparison to the level of involvement in Ellis II, and thus, the fact that higher-level personnel approved lower-level managers' recommendations, standing alone, was insufficient to establish commonality in a system of discretion. See id.

ii. Application to Plaintiffs' Disparate Impact Claims

As evident from the above survey, whether commonality exists in a system of discretion is a holistic inquiry that is highly dependent on the facts before the court. See supra § III.B.1.a.i. Applying the law on commonality to the case at hand, the court concludes that Plaintiffs have not shown that their disparate impact claims depend upon a // common contention, the determination of which will resolve an issue that is central to the validity of each class member's claims. See Dukes, 564 U.S. at 350.

At the outset, the court notes that Plaintiffs, in their briefing, unequivocally predicate their challenge on the discretion allowed under Microsoft's Calibration Process—precisely the argument that Dukes rejected as "just the opposite of a uniform employment practice that would provide the commonality needed for a class action." See id. at 355. Plaintiffs repeatedly attack the "lack of standardization" in the Calibration Process, noting that the process was "inconsistent" in terms of who attended, how managers compared employees, and what information managers considered. (MCC at 6.) Plaintiffs criticize how "evaluators were free to weight criteria for pay and promotions" in ways that were unconstrained by Microsoft's job requirements (id.) and how Microsoft did not exercise sufficient oversight to ensure that managers applied "the same standards consistently with each other" (id. at 7). Indeed, even Dr. Ryan's critique of the Calibration Process boils down to Microsoft's lack of guidance, resulting in a lack of uniformity across the process—in other words, unfettered discretion. (See Ryan Rep. ¶¶ 18-20, 26-31, 35-36.)

At oral argument, Plaintiffs attempted to distance themselves from the system of discretion they describe in their briefing, arguing for the first time that the use of Stock Levels dictated an employee's pay, leaving no room for discretionary decision-making. Put differently, Plaintiffs argued that after the managers submit their performance // // ratings¹¹—which Plaintiffs concede involve discretion—the employee's Stock Level mandates a certain pay with no discretion involved. But this attempt to divorce pay determinations from discretion is unavailing. It is true that peers were grouped, and thus compared, to others in similar Stock Levels (MCC Docs. at 13), but within those groups, managers still exercised discretion in comparing peers, resulting in various pay determinations, (see, e.g., Resp. Docs. at 55-56). Plaintiffs point to no evidence, and the court is not aware of any in the record, that an employee's Stock Level eliminates discretion because the Stock Level unequivocally determines one's pay.¹²

11 Plaintiffs concede that there is no gender disparity in the performance ratings. (MCC at 7.)

12 Moreover, Plaintiff's reliance on Stock Levels for pay determinations ultimately circles back to discretion. Even if Stock Levels determine pay, Stock Levels themselves are determined by promotions. (See Muenchow Decl. ¶ 2 (describing how Microsoft promoted her to higher Stock Levels).) And Plaintiffs admit that promotion decisions, in turn, involve discretion exercised by lower-level managers. Thus, even if women are disproportionately assigned to lower Stock Levels, the disproportionality itself is the result of discretionary decisions.

Of course, the fact that Plaintiffs' challenge centers on discretion does not foreclose commonality. See Ellis II, 285 F.R.D. at 518. But Plaintiffs must show that there is some common mode of exercising discretion that pervades the entire company and ties the alleged reasons for all those individual decisions together. See, e.g., Dukes, 564 U.S. at 352. Plaintiffs broadly assert that Microsoft "uses common criteria in a uniform Calibration [P]rocess" but offer little analysis beyond labeling certain processes and criteria as "common."¹³ (See Reply at 4.) Considering the size of the purported class, the Calibration Process's role as a mere framework, the subjective criteria set by Microsoft, and the lack of upper management involvement, the court concludes that Plaintiffs fail to identify a specific employment practice supplying a common question sufficient to certify a class.

13 Indeed, in their motion for class certification, Plaintiffs dedicate less than three pages to this critical issue of commonality. (See MCC at 32-34.) The argument is devoid of any citation to the record and thus fails to identify any Microsoft policy more specific than the "Calibration [P]rocess." (See id. at 32.)

First, Plaintiffs seek to certify a class of more than 8,600 women in various offices across the United States. (MCC at 1, 32.) Although 8,600 is a far cry from the 1.5 million in Dukes, 564 U.S. at 342, Plaintiffs' class is still many times the size of the certified classes in Ellis II, 285 F.R.D. at 509 (700 members), Chen-Oster, 2018 WL 1609267, at *2 (about 2,000 members), and Brown, 785 F.3d at 910 (100 members). Moreover, there is no indication that Plaintiffs' proposed class is "centralized" or "localized." See Brown, 785 F.3d at 910, 916. Unlike Ellis II—where all class members fell into two "closely related" positions sharing "a uniform job description across the class," 285 F.R.D. at 509—or Chen-Oster—where class members fell into two positions, 2018 WL 1609267, at *2—Plaintiffs' proposed class covers a myriad of positions, ranging from software engineers to game designers to data scientists, with an even more varied set of responsibilities within those positions (Whittinghill Decl. ¶¶ 21-35; see also id., Ex. C). Indeed, employees in the Engineering and IT Operations Professions fell into more than 8,000 unique positions for the period of time Plaintiffs challenge. (Id. ¶ 36.) And unlike Brown, Plaintiffs' putative class members were not located in a single facility that shared common spaces. See 785 F.3d at 910. Rather, they are located all over the United States, and even those employees in the Redmond headquarters are scattered across several facilities. (See SAC ¶¶ 2, 22.) Thus, Plaintiffs' proposed class is more akin to the expansive Duke class than the circumscribed classes in Ellis II, Chen-Oster, and Brown.

Second, the procedures Microsoft dictates act more as a framework than as constraints on discretion. Plaintiffs are correct that broadly speaking, all putative class members are subject to the "same, uniform compensation and promotion process." (See Reply at 2.) That is, all employees, including the putative class members, are first evaluated by their direct managers, then placed in peer groups of similar Stock Levels, and discussed in a broader meeting of managers—whether in a calibration meeting or a people discussion—with the consensus recommendation rolled up the management chain. (See id. at 2-3; Resp. Docs. at 39, 47-50.) But simply pointing out that Microsoft "set[] up a structure" through which all employees are evaluated is insufficient. See Valerino, 283 F.R.D. at 313. Like the evaluation and promotion process in Valerino, the process identified by Plaintiffs reveals only various stages or phases, during which individual decision-makers exercised their discretion. See id. at 313-14. Identifying the framework says nothing about whether the discretion itself can "be said to be the same or 'common.'" See id. at 313.

And unlike in Ellis II, where the company imposed requirements on how decisions were made, 285 F.R.D. at 511, the process by which Microsoft managers made decisions was largely left up to those managers. As Plaintiffs point out, it is true they had to hold a meeting. (See Resp. Docs. at 52-53.) But how they ran that meeting, and thus how they came to their decisions, was entirely discretionary. (See Helf Decl. ¶ 6; Wilson Decl. ¶ 15.) Each meeting leader determined how the discussion would be held, resulting in a "great deal of flexibility around the process and administration of the meeting." (Parris Decl. ¶ 6, Ex. 4 ("Coleman Dep.") at 31:19-21.) One group may decide to use physical, color-coded cards to represent each employee. (DeCaprio Decl. ¶ 9.) Another group may utilize a PowerPoint presentation to visually display relevant rating and budgetary information. (Id.) Still another group may dispense with visual props altogether, launching instead into a roundtable discussion. (Id. ¶ 10.) Some groups may discuss clusters of similarly-rated employees at a time; others may only focus on the high and low performers for each manager. (See id. ¶ 14.) Some groups may require consensus amongst the attending managers; others may not. (Id. ¶ 15.)

Thus, "increasing the level of abstraction" to the Calibration Process does not alter the flexibility granted to local managers, and nothing about the Calibration Process itself provides the "common direction" necessary to establish that individual managers exercised their discretion in a common way. See Jones, 34 F. Supp. 3d at 904. That there was some structure to the Calibration Process "does not change the fact that the structure reinforced the discretionary nature of the decisionmaking in this area." See id. at 906. Plaintiffs seem to acknowledge as much, as they challenge the "lack of standardization" in "the specific procedures for discussing and making compensation and promotion decisions." (MCC at 6.)

Third, Microsoft's general criteria do not sufficiently constrain discretion. Quite the opposite. Microsoft's benchmarks are subjective and open to many interpretations— indeed, Microsoft designed the criteria so that managers could adapt it to various employees' roles. For the calibration meetings, Microsoft offered general definitions of the "what," "how," and "proven capability" criteria, as well as the 1-5 numerical rating system. For instance, Microsoft defined the "how" input as "[h]ow results were accomplished for the past fiscal year." (Resp. Docs. at 55.) But Microsoft specifies that the "degree to which each input factors into the overall assessment will likely vary by role" and that the "[o]verall performance and ratings reflect the environment in which results were achieved," all of which necessarily differs from team to team. (Id. at 50-51; see also Helf Decl. ¶ 16 ("[T]he criteria used to evaluate individuals' performance within [different] groups were necessarily different.").) And even within this general criteria, managers were encouraged to come to their own understanding of "common objectives" and "what good work looked like," which effectively allowed managers to consider any other criteria as they saw fit. (See DeCaprio Decl. ¶ 8.) Thus, like the "communication skills" or "ability to learn" criteria in Dukes II, Microsoft's subjective criteria were so vague "that they imposed no real constraints."¹⁴ See 964 F. Supp. 2d at 1126.

14 When asked during oral argument what evidence Plaintiffs had that the "what," "how," and "proven capacity" criteria were uniformly understood by lower managers, Plaintiffs pointed to Microsoft's training on the criteria. But Plaintiffs' reliance on Microsoft's training belies their own expert's evaluation: Dr. Ryan chastised Microsoft's training as inadequate to provide sufficient guidance. (See Ryan Rep. ¶ 43.) Plaintiffs cannot have it both ways.

After Microsoft transitioned from calibration meetings to people discussions, its evaluation criteria became even less concrete: Managers evaluate "the impact [a] particular employee's performance had on the group and the company." (See Wilson // Decl. ¶ 23.) Like the "amorphous concepts" in Jones, 34 F. Supp. 3d at 906, it is hard to imagine a criterion more susceptible to individual interpretation than the subjective measure of a person's "impact." Unsurprisingly then, managers in people discussions "ha[d] much more latitude" in evaluating an employee's impact. (Helf Decl. ¶ 35.) Indeed, "any metric can be considered, so long as it is relevant to the employees' impact." (Id.) Such wide and varied considerations can hardly be said to be commonly understood by all employees. See Ellis II, 285 F.R.D. at 514. Again, Plaintiffs seem to recognize that reality, arguing that "decision makers were able to apply variable standards in making compensation and promotion decisions." (MCC at 6.)

Lastly, Plaintiffs have not shown sufficient involvement by top management. Plaintiffs identify the four EVPs as the final approvers of all pay and promotion decisions and correctly observe that the performance rating recommendations rolled up the ranks. (Reply at 2-3, 12.) But the evidence shows that the upper-level executives "almost always accepted" the recommendations of lower-level managers. See Jones, 34 F. Supp. 3d at 908. Microsoft senior management rarely revisited the recommendations made at the discretion of lower-level managers; indeed, it would be "highly unusual" for senior management "to make a change to a rewards recommendation." (Helf Decl. ¶ 36; see also Shepherd Decl. (Dkt. # 315) ¶ 13 ("[T]he partner-level managers typically approved the individual recommendations made at the lower-level [c]alibration [m]eetings."); DeCaprio Decl. ¶ 16 (observing that senior leadership will not closely review recommendations for levels below Level 68); Jarvis Decl. (Dkt. # 306) ¶¶ 17-18 (noting // that there is "typically little movement for recommendations affecting employees in Level 67 roles and below" by the time the process rolls up to the EVP).)

Thus, although EVPs approve pay and promotion decisions, their involvement is more akin to the "limited oversight" in Jones, see 34 F. Supp. 3d at 908, than the "consistent[] and pervasive" involvement in Ellis II, see 285 F.R.D. at 511-12. As in Jones, Plaintiffs provide almost no evidence that the EVPs changed pay or promotion recommendations with "any frequency," much less that they did so with "regularity."¹⁵ See 34 F. Supp. 3d at 908. There is additionally no evidence, as there was in Ellis II, that the EVPs "actually make the decisions," see 285 F.R.D. at 515; the EVPs did not pick the eligible candidates for promotion, conduct floor-walks to assess potential applicants, or personally instruct lower-level managers on how to execute the Calibration Process,¹⁶ see id. at 512-14. Accordingly, Plaintiffs' evidence of the EVPs as a core group who influenced the discretionary decisions falls short.

15 In support of this point, Plaintiffs at oral argument identified only one statement by Senior HR Manager Shannon Shepherd that she has "seen . . . executive vice presidents modify promotion recommendations" for budgetary reasons. (Shepherd Decl. ¶ 25.) There is no indication that those modifications are regular practice. (See id.)

16 Plaintiffs conceded at oral argument that the role upper management played in Ellis II was qualitatively different from the role Microsoft's upper management played in pay and promotion determinations here.

On this issue of whether Plaintiffs have identified a common mode of exercising discretion that pervades Microsoft, Plaintiffs' expert reports are of no assistance. As discussed above, Dr. Ryan, rather than identifying sufficient common discretion, emphasizes the exact opposite: the "lack of standardization in how factors are evaluated // and the lack of standardization in the process itself." (See Ryan Rep. ¶ 25; see also Parris Decl. ¶ 15, Ex. 13 ("Ryan Dep.") at 209:2-213:17 (focusing critique on the managers' freedom to apply and weigh criteria differently).) Indeed, Dr. Ryan's report resembles that of the plaintiffs' expert in Dukes, who could not calculate "whether 0.5 percent or 95 percent of the employment decisions . . . might be determined by stereotyped thinking." See 564 U.S. at 354. Such testimony, as the Supreme Court found, could be "safely disregard[ed]." Id.

Likewise, Dr. Farber's aggregate statistical analysis, which purports to establish the disparate impact felt by female employees, sheds no light on whether Microsoft had a company-wide policy constraining the discretion of lower-level managers. (See Parris Decl. ¶ 9, Ex. 7 ("Farber Dep.") at 264:13-23 (stating that analysis did not touch on the Calibration Process's role in any identified disparities).) "[S]tatistical correlation cannot substitute for a specific finding of class-action commonality." Campbell v. Nat'l R.R. Passenger Corp., --- F. Supp. 3d ----, 2018 WL 1997254, at *31 (D.D.C. Apr. 26, 2018); see Dukes, 564 U.S. at 357 ("Merely showing that Wal-Mart's policy of discretion has produced an overall sex-based disparity does not suffice."). Thus, as in Dukes, even if Dr. Farber's report conclusively establishes disparities based on gender, "that would still not demonstrate that commonality of issue exists." See 564 U.S. at 357.

Indeed, Plaintiffs' proffered declarations reflect how Microsoft managers did not exercise their discretion in a uniform manner. The declarants concede that, although some managers exercised discretion in a discriminatory manner, others did not. (See, e.g., Parris Decl. ¶ 18, Ex. 16 ("Underwood Dep.") at 198:19-199:13 (agreeing that whether a manager discriminated was "event-specific and person-specific" because the choice to discriminate is specific "to that person themselves [sic]").) Moreover, the declarants' experiences varied and were based on a number of different policies or practices. For example, one declarant spoke of discrimination she saw from the recruitment department when she was hiring candidates. (Boeh Decl. ¶ 6.) Another described alleged retaliation from her direct supervisor after she complained to HR. (Smith Decl. ¶ 10.) Thus, Plaintiffs' declarations further illustrate how employees experienced discrimination in different ways at the hands of different individuals.

The cases that Plaintiffs highlight in which courts certified Title VII classes all feature some "glue" that tied the many discretionary decisions together—the very thing that is missing here. As the court has already discussed at length, several features distinguish the case-at-hand from Ellis II. Unlike Ellis II, the class size here is not as limited in number or scope, see id. at 509; no companywide process imposes meaningful requirements on lower-level managers, see id. at 511; and critically, top management is not intimately involved with the pay and promotions process, see id. at 511-15.¹⁷ // //

17 Likewise, the proposed class here is distinguishable from that in Chen-Oster. The proposed class in Chen-Oster was four times smaller than the class here and spanned only two job positions. See 2018 WL 1609267, at *2. Moreover, Chen-Oster featured more involvement by upper-management, such as setting the budget that determined compensation (id. at *4), developing a list of candidates for promotion with ranking to emphasize priority candidates (id.), and personally selecting and training the people who evaluated the promotion candidates (id. at *5). Ultimately, the upper management "decide[d] who is promoted." (Id.) There is no such evidence of upper management involvement here.

And unlike McReynolds, Plaintiffs do not identify a companywide policy that required managers to allow the allegedly-discriminatory practice. See 672 F.3d at 489-90. In McReynolds, plaintiffs challenged "teaming" and "account distribution" policies that allowed employees to form their own teams for compensation purposes. Id. at 488. But these policies did not allow any discretion on the part of managers in determining whether to allow teaming. Id. at 489. Indeed, McReynolds recognized that if the decision of whether to allow teaming was delegated to local management, the resulting discrimination would be akin to that alleged in Dukes. Id. at 490. Because the challenged policies were mandated on a company-wide basis, however, the policies were not an exercise of discretion by local managers. Id. at 489-90. Plaintiffs identify no comparable corporate policy here.¹⁸

18 The last case Plaintiffs identified also involved a mandatory system that left no room for managerial discretion. (See MCC at 33 (citing Parra v. Bashas', Inc., 291 F.R.D. 360, 375 (D. Ariz. 2013)).) In Parra, plaintiffs challenged the company's use of two wage scales that resulted in different pay for employees of different races. 291 F.R.D. at 373-74. The court explicitly recognized that the wage scales were "non-discretionary"—that is, local managers did not have any choice in implementing the two wage scales and the corresponding pay levels. Id. at 375. By contrast, Microsoft's Calibration Process left significant choice to the lower-level managers in determining how to arrive at pay and promotion decisions.

In sum, Plaintiffs challenge Microsoft's policy of allowing discretion by lower-level managers but have not identified a common mode of exercising discretion that pervades the entire company. As in Dukes, without some common direction, it is "quite unbelievable" that all Microsoft managers supervising over 8,600 putative class members "would exercise their discretion in a common way." 564 U.S. at 356. Thus, Plaintiffs are in the same position as the Dukes plaintiffs: "challenging [a] practice of // delegating discretion to local managers, which the Supreme Court specifically held was not a specific employment practice supplying a common question sufficient to certify a class." See Dukes II, 964 F. Supp. 2d at 1127. Because Plaintiffs provide no convincing evidence of "some glue" holding together the reasons behind the numerous employment decisions they challenge, they have not established a common answer to the question "why was I disfavored." See Dukes, 564 U.S. at 351. Accordingly, the court concludes that commonality is lacking for Plaintiffs' disparate impact claims.

b. Disparate Treatment Commonality

Unlike disparate impact claims, disparate treatment claims do not require plaintiffs to identify a specific companywide employment practice responsible for the discrimination. See Int'l Bhd. of Teamsters v. United States, 431 U.S. 324, 336 n.16 (1977). Instead, plaintiffs must provide evidence of a "systemwide pattern or practice" of discrimination, such that the discrimination is "the regular rather than the unusual practice." Id. at 336. This burden may be met through statistics alone. Id. at 339-40. Upon such a showing, the required discriminatory intent may be inferred. See id. If plaintiffs establish a prima facie case, the burden shifts to the employer to show that the plaintiffs' statistical evidence is either "inaccurate or insignificant." Id. at 360.

There is a "wide gap" between an individual's claim that he was subjected to a company's policy of discrimination, and the existence of a class of persons who have suffered the same injury as that individual, such that the individual's claim and the class claims will share common questions of law or fact. Dukes, 564 U.S. at 352-53 (citing Falcon, 457 U.S. at 157-58). To bridge that gap, Plaintiffs must provide "significant proof" that Microsoft "operated under a general policy of discrimination." Id. at 353. Here, Plaintiffs point to (1) statistical evidence of gender-based disparities; (2) anecdotal evidence illustrating female employees' experiences of discrimination at Microsoft; and (3) documents evincing Microsoft's culture of hostility towards women. (See Reply at 14.) The court finds, however, that none of the proffered categories of evidence constitute the necessary "significant proof."¹⁹

19 As a preliminary matter, the court notes that much of the commonality analysis for Plaintiffs' disparate impact claims apply with equal force to the disparate treatment claims. See Ellis II, 285 F.R.D. at 511 (recognizing the identification of "common policies and practices" as a category of proof supporting commonality in disparate treatment claims). Because Plaintiffs have not identified the existence of common pay and promotion practices that apply to the class as a whole, see supra § III.B.1.a, they cannot predicate commonality for their disparate treatment claims on a "common, companywide promotion system . . . made up of numerous common components," as the plaintiffs in Ellis II did. See id. at 511-19.

i. Statistical Evidence

The court first addresses Plaintiffs' statistical evidence conducted by Dr. Farber. The plaintiffs in Dukes offered a similar statistical study, in which regression analyses revealed statistically significant disparities between men and women that "can be explained only by gender discrimination." 564 U.S. at 356. Dukes held that this statistical evidence was insufficient because it suffered from a "failure of inference." Id. The evidence established only that there was a disparity at the national or regional level, but information at that level did not establish the existence of disparities at individual stores, let alone "the inference that a company-wide policy of discrimination is implemented by discretionary decisions at the store and district level." Id. at 356-57. // // Thus, these statistics could not produce a common answer to the question "why was I disfavored." Id. at 352.

Post-Dukes, the statistical study must "conform[] to the level of decision for the challenged practices." Ellis II, 285 F.R.D. at 523. When decisions are made at the discretion of each individual supervisor, courts generally find aggregated statistical evidence inadequate because it is "derived from hundreds of employment decisions made by myriad decision makers, at different times, under mutable procedures and guidelines, in different departments, and in different office locations, concerning employees at varying levels of experience, responsibilities, and education." Jones, 34 F. Supp. 3d at 909. On the other hand, if plaintiffs demonstrate that practices are uniform across the company, then courts have found "good reason to rely on nationwide statistics." Ellis II, 285 F.R.D. at 523 (relying on aggregated statistics because the CEO averred that the promotion policies derived from the top management and were uniform across the company).

Plaintiffs here have not shown, as the plaintiffs in Ellis II did, that promotion policies and practices are uniform across Microsoft. Compare supra § III.B.1.a, with Ellis II, 285 F.R.D. at 511-18. Thus, unlike in Ellis II, there is not the same "good reason to rely on nationwide statistics." See 285 F.R.D. at 523. The relevant level of decision-making for the challenged practices here remains at the individual manager, or at best, the team level. Dr. Farber's statistical study, however, is based on aggregate figures, including all individuals in the Engineering or IT Operations Professions in Stock Levels 59-67. (Farber Rep. ¶ 50 n.45.) At most, he disaggregates the data no further than the level of the four EVPs, whom the court has already determined merely approved the pay and promotion decisions. (Farber Rebuttal (Dkt. ## 344 (sealed), 359-10 (redacted)) ¶ 16); see supra § III.B.1.a. Therefore, even accepting the validity of Dr. Farber's statistical analysis, his evidence "has the same problem as the statistical evidence in [Dukes]: it begs the question." See Bolden, 688 F.3d at 896. Dr. Farber's finding of disparity "may be attributable to only a small set of [Microsoft managers], and cannot by itself establish the uniform, [manager-by-manager] disparity upon which the [P]laintiffs' theory of commonality depends."²⁰ See 564 U.S. at 357. Consequently, Dr. Farber's statistical evidence is insufficient to show any common issue that would permit a nationwide class.²¹

20 Bolden provides an example to illustrate this effect. See 688 F.3d at 896. For instance, if Microsoft had 25 managers, 5 of whom discriminated in making pay and promotion decisions, aggregate data would show that female employees fared worse than male employees. But "that result would not imply that all 25 [managers] behaved similarly, so it would not demonstrate commonality." Id.

21 Moreover, Dr. Farber's analysis does not link the gender disparity he found to the challenged Calibration Process. (See Farber Rep.; Farber Dep. at 264:13-24.) Thus, in comparison to Chen-Oster, where Dr. Farber concluded that the disparity stemmed from the very corporate policies being challenged, Dr. Farber's analysis here does not as readily support an inference of a policy of discrimination. See 2018 WL 1609267, at *15.

ii. Anecdotal Evidence

Plaintiffs' anecdotal evidence also do not constitute the necessary "substantial proof." In Dukes, the Supreme Court found the proffered anecdotal evidence "too weak" because plaintiffs filed only 1 affidavit for every 12,500 class members—a ratio that paled in comparison to the 1 for every 8 class members proportion the Court accepted in Teamsters v. United States, 431 U.S. at 337-38. Dukes, 564 U.S. at 358. Courts following Dukes have similarly required a significant amount of anecdotal evidence. See Brown, 785 F.3d at 913 (1 for every 6.25 class members); Rollins, 2016 WL 258523, at *5, 7-8 (about 1 for every 4 class members). Here, by contrast, Plaintiffs offer 9 declarations for 8,630 class members, or about 1 for every 959 class members.²² (See Farber Rep. ¶ 11; see generally Dkt.) Thus, on the numbers alone, Plaintiffs' anecdotal evidence is "too weak" to establish that the entire company operates under a general policy of discrimination. See Dukes, 564 U.S. at 358 n.9 ("[W]hen the claim is that a company operates under a general policy of discrimination, a few anecdotes selected from literally millions of employment decisions prove nothing at all.").

22 Even including the named plaintiffs' declarations, Plaintiffs' total rises to 11 declarations for 8,630 class members, or about 1 for every 785 class members. (See Farber Rep. ¶ 11; see generally Dkt.)

Moreover, the anecdotal evidence relates to only 5 of the 41 states in which Microsoft has offices. (See generally Miller Decl.; Underwood Decl.; Hutson Decl.) For four of those five states, Plaintiffs provide only one declaration. (See generally id.) Nor do the declarations represent all the Stock Levels in the class. There are no declarations from Level 67 employees, only one declaration from a Level 66 employee, and only three from Level 59 employees. And lastly, it is difficult to imagine that these nine declarants represent sufficiently the multitude of positions throughout the company. See supra § III.B.1.a.ii (noting over 8,000 unique positions in the challenged Professions).

The court does not, in any way, seek to minimize the gravity of these employees' experiences. To the contrary, they provide many examples of serious misconduct. (See, // e.g., Boeh Decl. ¶ 5 (stating that a male manager refused to recommend her for a promotion because of her potential to take maternity leave).) But even taking all the accounts as true, the nine accounts are simply not enough to demonstrate that Microsoft operated under a general policy of discrimination towards over 8,600 female employees across 41 states holding thousands of unique positions.²³

23 Microsoft additionally argues that the declarations "suffer from many . . . evidentiary problems." (Resp. at 31 n.21.) But the Ninth Circuit recently held that "[i]nadmissibility alone is not a proper basis to reject evidence submitted in support of class certification." Sali, 2018 WL 2049680, at *5. Because the court's consideration "should not be limited to only admissible evidence," Microsoft's argument regarding the admissibility of Plaintiffs' declarations is misplaced. See id.

iii. Culture Evidence

Plaintiffs next assert that the declarations need to be considered in combination with a third category of evidence: documents, such as internal Microsoft memoranda and employee correspondence, that exhibit Microsoft's culture of hostility towards women. (See Reply at 15; see also MCC at 13-27.) These documents include employee reactions to Microsoft's announcements that it had achieved gender pay equity (MCC at 13-16); senior management's knowledge of gender bias (MCC at 16-18); and various complaints and charges regarding gender discrimination (MCC at 19-28).

But, viewed either individually or in combination, this evidence suffers from the same problem as the declarations: the evidence does not rise to the level of a general, company-wide policy of discrimination that is tied to the challenged employment decisions. Dukes, again, is illustrative. The Dukes plaintiffs similarly relied on evidence of Wal-Mart's "strong corporate culture," which purportedly made the company vulnerable to gender bias. 564 U.S. at 353-54. However, plaintiffs did not show "with any specificity how regularly stereotypes play a meaningful role in employment decisions" within the alleged culture of bias. Id. at 354. Without such evidence, information regarding Wal-Mart's culture did "nothing to advance [the plaintiffs'] case," because the regularity with which employment decisions might be determined by biased thinking "is the essential question on which [plaintiffs'] theory of commonality depends." Id.

Here, Plaintiffs' evidence offer, at most, a glimpse into individual incidents that are not sufficiently representative of the entire culture across Microsoft. For instance, Plaintiffs rely on employee reactions to two Microsoft announcements regarding its progress in pay equity. (See, e.g., MCC Docs. at 516-17.) But these reactions—while they may be legitimate critiques of Microsoft's pay equity analysis—reflect only those employees' belief regarding the pay equity issue; such personal beliefs fall short of illustrating a common corporate culture. Indeed, Plaintiffs' evidence reveals that Microsoft CEO Satya Nadella, as well as others, responded to the critiques by acknowledging the next steps Microsoft must take to further address the issue. (E.g., id. at 516.) Such a response does not reflect what Plaintiffs claim to be a culture of evasion and refusal to acknowledge the problem.²⁴ (See MCC at 16.) //

24 For this reason, the court also finds unpersuasive Plaintiffs' evidence regarding the SLT's recognition of Microsoft's areas of growth when it comes to D&I matters. (See MCC at 16-18.) Apologizing for past missteps, acknowledging areas of improvement, and setting a "Priorities and Action Plan" do not reflect inaction. (See, e.g., MCC Docs. at 38-39 (recognizing three areas that Microsoft must immediately begin work on).)

The same goes for Plaintiffs' evidence of various gender discrimination complaints. Plaintiffs assert that the 238 internal complaints Microsoft received are "shocking" but provide the court with no evidence regarding whether that number is unusual for a company like Microsoft with hundreds of thousands of employees. (See MCC at 24.) Plaintiffs additionally attack the efficacy of Microsoft's ERIT but, again, provide the court no official evaluation of the team other than employee commentary. (See id. at 24-25.) And lastly, Plaintiffs point to a Notice of Violation issued by the Department of Labor's OFCCP. (See MCC at 11-13; see also MCC Docs. at 112-24.) But the OFCCP had conducted several other audits of Microsoft offices and found no violations on at least 18 other occasions. (See Parris Decl. ¶ 21, Ex. 19.) Thus, the evidence regarding the complaints does not satisfy the "significant proof" needed at this stage.²⁵

25 Even accepting the veracity of Plaintiffs' claims about Microsoft's culture, this evidence is insufficient to establish commonality. Like the plaintiffs' culture evidence in Dukes, Plaintiffs here do not indicate "with any specificity" how the alleged gender stereotypes stemming from Microsoft's culture play a meaningful role in employment decisions. See 564 U.S. at 354. Without answering this "essential question" upon which commonality depends, the evidence of Microsoft's corporate culture is insufficient to establish commonality. See id.

On the whole, the court finds that Plaintiffs have not provided the "significant proof of Microsoft's "general policy of discrimination" necessary to demonstrate commonality in their disparate treatment claims. See Dukes, 564 U.S. at 353. The proffered statistical evidence, the anecdotal evidence, and the internal Microsoft documents do not establish the existence of any common issue, and accordingly, Plaintiffs have not satisfied commonality for their disparate treatment claims. Because commonality is lacking for both Plaintiffs' disparate impact and disparate treatment claims, Plaintiffs have failed to satisfy Rule 23(a)'s prerequisites.²⁶

26 Because Plaintiffs have failed to satisfy commonality, they also have failed to show the "far more demanding" requirement that common questions predominate over any questions affecting only individual members. See Amchem Prods. Inc. v. Windsor, 521 U.S. 591, 624 (1997); Fed. R. Civ. P. 23(b)(3); see also Hanlon v. Chrysler Corp., 150 F.3d 1011, 1022 (9th Cir. 1998) (noting that the Rule 23(b)(3) analysis "presumes that the existence of common issues of fact or law have been established . . . . [T]he presence of commonality alone is not sufficient to fulfill Rule 23(b)(3)").

2. Typicality

Rule 23(a)(3) requires that "the claims or defenses of the representative parties are typical of the claims or defenses of the class." Fed. R. Civ. P. 23(a)(3). "The purpose of the typicality requirement is to assure that the interest of the named representative aligns with the interest of the class." Hanon v. Dataproducts Corp., 976 F.2d 497, 508 (9th Cir. 1992). Courts "do not insist that the named plaintiffs' injuries be identical with those of the other class members." Armstrong v. Davis, 275 F.3d 849, 869 (9th Cir. 2001). Instead, in determining typicality, the court asks "whether other members have the same or similar injury, whether the action is based on conduct which is not unique to the named plaintiffs, and whether other class members have been injured by the same course of conduct." Ellis I, 657 F.3d at 984 (internal quotation marks omitted) (quoting Hanon, 976 F.2d at 508). Individualized defenses applicable to the class representatives do not preclude a finding of typicality unless there is a danger that absent class members will suffer if their representative is preoccupied with defenses unique to it. Hanon, 976 F.2d at 508. //

The "commonality and typicality requirements of Rule 23(a) tend to merge." Falcon, 457 U.S. at 157 n.13; see Parsons v. Ryan, 754 F.3d 657, 685 (9th Cir. 2014). "Where the challenged conduct is a policy or practice that affects all class members, the underlying issue presented with respect to typicality is similar to that presented with respect to commonality, although the emphasis may be different." Armstrong, 275 F.3d at 868-69. Thus, "if the plaintiffs cannot establish that they were injured by the same conduct that injured other class members, then their claims cannot be typical of other members of the class." Jones, 34 F. Supp. 3d at 911; see also Donaldson v. Microsoft Corp., 205 F.R.D. 558, 568 (W.D. Wash. 2001) ("[W]here there is no evidence of a common experience shared by all [class members] at Microsoft, there can also be no 'typical' class representative.").

Plaintiffs contend that Ms. Moussouris and Ms. Muenchow's claims are identical to those of all other class members because they "[have] been paid less than comparable male coworkers" and were passed over "for promotions in favor of less qualified and less experienced men." (MCC at 35.) The court concludes otherwise and finds, consistent with its commonality determinations, that Plaintiffs have not established typicality with respect to their disparate impact and disparate treatment claims.

Although Plaintiffs have identified that all putative class members would be challenging the same injury—lower pay or missed promotions—Plaintiffs' generalized characterization obscures the many unique facets of the underlying conduct, stemming from the discretion exercised by countless lower-level managers. See supra § III.B.1.a.i. In other words, Plaintiffs have not shown that "other class members have been injured by the same course of conduct." See Ellis I, 657 F.3d at 984. Some individual class members may want to argue that their initial performance rating was flawed. Or some may want to argue that their accomplishments were not adequately represented in the calibration meeting or people discussion. Others may have substantial evidence that their particular manager was biased. And still others may have no qualms with their initial performance rating, the first calibration meeting, or their direct manager, but wish to challenge how successive discussions unfolded in the roll-up process.

Indeed, Plaintiffs' proffered declarations illustrate the many differences among individual class members' challenges. For instance, Ms. Dove may challenge the forced comparison process in calibration meetings, because "someone had to receive zero rewards, and that someone became [her] because of [her] gender." (Dove Decl. ¶ 5.) Ms. Dove's discrimination claim, then, would center on those calibration meetings, the managers present, and the specific peer group to which she was compared. See id. Ms. Boeh, on the other hand, would challenge her manager's initial recommendation—or lack thereof—when the manager "did not want to 'waste' a promotion" on her after her return from maternity leave. (Boeh Decl. ¶ 5.) Thus, Ms. Boeh's discrimination claim would not concern the forced comparison in calibration meetings but instead, the initial recommendation process and the specific manager who refused to recommend her for a promotion. Both claims may well have merit, but the very nature of these claims—that is, the conduct that injured these declarants—differs from that of Plaintiffs' claims and from each other. See Jones, 34 F. Supp. 3d at 911. //

In short, the court agrees with the Valerino court's analysis on how a discretionary pay and promotion system defeated typicality:

This is all to say nothing of the different facts plaintiffs would rely on to prove their disparate claims—different ranking lists, different [manager] notes, different characteristics of individuals like years on the job, district assignments, and collateral duties, different number of applicant applications, etc. . . . And it is clear that [Plaintiffs'] claims at least involve different [managers]. This alone is sufficient to demonstrate that their claims are not typical of other members of the class.

283 F.R.D. at 318. The conduct underlying Plaintiffs' action necessarily differs from the conduct that caused other class members' injuries in this system of unrestrained discretion. Given the discretion-based system, the individualized inquiry will vary for each plaintiff, foreclosing any contention that the named plaintiffs' claims are typical. Thus, the court concludes that Plaintiffs have also failed to establish typicality.

3. Adequacy

Rule 23(a)(4) requires Plaintiffs to demonstrate that they will fairly and adequately protect the interests of their class. Fed. R. Civ. P. 23(a)(4). To determine adequacy, the court must resolve two questions: "(1) do the named plaintiffs and their counsel have any conflicts of interest with other class members[;] and (2) will the named plaintiffs and their counsel prosecute the action vigorously on behalf of the class?" Hanlon, 150 F.3d at 1020.

Microsoft does not dispute the second issue—that Plaintiffs and counsel will prosecute the action vigorously—nor is the court aware of any reason to doubt Plaintiffs and their counsel on this point. (See Resp.) Microsoft instead argues that adequacy is lacking because the proposed class presents irreconcilable conflicts of interest. (See id. at 39-41.) Specifically, Microsoft maintains that because the proposed class consists of females employees who participated as managers in the Calibration Process—the very process being challenged—Plaintiffs must "impugn[] the input of thousands of class members who participated in calibration." (Id. at 41.) The court agrees.

Although there is no per se rule regarding adequacy where a class includes "employees at different levels of an employment hierarchy," "[the] concern about classes that involve both supervisors and rank-and-file workers can be a valid one in some circumstances." Staton v. Boeing Co., 327 F.3d 938, 958 (9th Cir. 2003). Courts have not only held that "supervisors may not be appropriate representatives of their subordinates" but have also extended that logic "to prevent a subordinate from representing a supervisor." Pena v. Taylor Farms Pac., Inc., 305 F.R.D. 197, 215 (E.D. Cal. 2015). "[W]hether employees at different levels of the internal hierarchy have potentially conflicting interests is context-specific and depends upon the particular claims alleged in a case." Staton, 327 F.3d at 958.

Donaldson v. Microsoft Corp. presented circumstances where such a conflict existed. See 205 F.R.D. at 568. In that case, the court expressed "serious concerns" because two named plaintiffs were former supervisors who "were obligated to implement the very supervisory system which this litigation challenges." Id. Moreover, class members were potentially "in conflict with each other" because of the "significant number of potential class members who are current or former managers." Id. Because the allegations of disparate impact and disparate treatment arose "directly from the evaluation system," the court was "unable to envision a class which would include both those who implemented the ratings system and those who allegedly suffered under it." Id.; see also Valerino, 283 F.R.D. at 318 (recognizing potential conflicts amongst class members if two putative class members were seeking the same promotion).

By contrast, in Pena v. Taylor Farms Pacific, Inc., the fact that the class included both supervisors and their subordinates did not defeat adequacy. See 305 F.R.D. at 215. The class in that case alleged that the company implemented uniform policies to withhold pay, deny meal and rest breaks, and issue erroneous paychecks. Id. at 203-04, 215. The court observed that if these uniform policies required hourly employees to work without compensation and breaks, "both supervisors and subordinates would have been affected similarly" because both were hourly employees. Id. at 215. The class member's title therefore had no relevance to the challenged policies; supervisor or subordinate, the class member would have been denied compensation and break time. See id. Additionally, there was no evidence that the supervisors would be liable for the alleged violations—that is, they had not implemented the policies. Id. Thus, the court concluded that "[n]o conflict is sufficient to call adequacy into question." Id.

The case at hand more closely resembles Donaldson. As in Donaldson, Ms. Moussouris, a named plaintiff, participated in the Calibration Process as a manager and thus was "obligated to implement" the very system that Plaintiffs challenge. (See Saad Rep. ¶ 18; Moussouris Dep. at 103:16-23, 104:19-21.); 205 F.R.D. at 568. Indeed, Ms. Moussouris recalls some of her subordinates' ratings being downgraded during the Calibration Process. (Moussouris Dep. at 119:4-18.) Furthermore, 2,126 putative class members were managers at least once during the class period, and 3,457 members were either leads or managers; 472 qualified as "managers of managers."²⁷ (Saad Rep. ¶ 17.) Thus, as in Donaldson, a "significant number of potential class members" here participated in the very system and the very decisions that are alleged to be discriminatory. See 205 F.R.D. at 568. The direct involvement of the managers, and the extent of that involvement, distinguish the circumstances here from those in Pena, where the supervisors had no say in the challenged policies. See 305 F.R.D. at 215. Because Plaintiffs' disparate impact and disparate treatment claims stem directly from the system that many putative class members—including a named plaintiff—took part in, the class "include[s] both those who implemented the ratings system and those who allegedly suffered under it." See Donaldson, 205 F.R.D. at 568. The court agrees with Donaldson that "[t]his conflict appears insurmountable." See id.

27 At least three of the nine declarants also participated in the Calibration Process as managers. (See Parris Decl. ¶ 4, Ex. 2 ("Alberts Dep.") at 69:14-18; id. ¶ 8, Ex. 6 ("Dove Dep.") at 44:3-22; id. ¶ 15, Ex. 15 ("Sowinska Dep.") at 54:22-55:8.) --------

Plaintiffs' only response to Donaldson is that the Ninth Circuit's subsequent Staton v. Boeing Co. decision limited Donaldson's applicability. (See Reply at 20.) Not so. Staton did not abrogate Donaldson or call into doubt Donaldson's reasoning. Staton, 327 F.3d at 958. Instead, Staton merely affirmed that for the purposes of adequacy, the significance of a class with employees at different levels depends on the circumstances of the case. Id. Like the circumstances presented in Donaldson, 205 F.R.D. at 568, and contrary to those in Pena, 305 F.R.D. at 215, the context and claims presented here lead // // the court to conclude that Plaintiffs and absent class members' interests will often conflict. Those conflicts of interest further undermine class certification.

In summary, after careful review, the court finds that Plaintiffs have failed to carry their burden of satisfying the Rule 23(a) prerequisites. First, Plaintiffs have not demonstrated a common question to be resolved on behalf of the putative class. See Fed. R. Civ. P. 23(a)(2). Additionally, Plaintiffs have failed to establish that the Plaintiffs' claims are typical of those of the class members, see id. 23(a)(3), or that the Plaintiffs are adequate representatives of absent class members, see id. 23(a)(4). For all of these reasons, the court denies Plaintiffs' motion for class certification.

IV. CONCLUSION

For the foregoing reasons, the court DENIES Plaintiffs' motion for class certification (Dkt. ## 228 (sealed), 381 (redacted)). The court DIRECTS the Clerk to provisionally file this order under seal and ORDERS the parties to meet and confer regarding the need for redaction. The court further ORDERS the parties to jointly file a statement within ten (10) days of the date of this order to indicate any need for redaction.

Dated this 25th day of June, 2018.

/s/_________

JAMES L. ROBART

United States District Judge

³ [Redacted]