The Industry

Google Revealed It Was Underpaying Some Men. That’s Misleading—and Damaging.

A man rides a bike past a Google sign at the Googleplex in Menlo Park, California.
The Googleplex.
Josh Edelson/Getty Images

Google says it’s important to pay employees based on what they deserve, not based on who they are. But as decades of employment data demonstrate, that just doesn’t happen, even at places that regard themselves as exceptionally meritocratic, like Google. To try to fix pay imbalances, for the past seven years Google has conducted what it calls an “annual pay equity analysis” in which the company examines the compensation of workers in various job groups and corrects any significant gaps across race and gender. On Monday, it shared a small handful of results from its 2018 analysis.

You probably heard about one shocking finding. In one job category (“Level 4 Software Engineers”), Google says it actually found it was underpaying men—a shortfall it then corrected by issuing raises to thousands of men across the company. Overall, pay bumps resulting from the analysis added up to $9.7 million for more than 10,600 employees, though the company didn’t say how many of those employees were men. But James Finberg, a lawyer for a class-action lawsuit filed against Google for underpaying women, told Wired that the compensation adjustment was granted to some 8,000 male engineers. Wired reports that the discrepancy resulted from managers allocating more of their discretionary budget to women. Google reportedly decided to share this particular detail from the report because it was “counterintuitive.”

Counterintuitive is one way to put it. Misleading is another. Google’s yearly analysis examines the pay of employees in “the same job, at the same level, location and performance,” the company’s lead analyst for pay equity, Lauren Barbato, wrote in a blog post. What the review doesn’t cover is whether women hired with the same levels of experience as men are given lower positions at the company—a problem called leveling. As it happens, that’s one issue that the class-action lawsuit representing more than 8,300 current and former Google employees alleges is part of the problem. One plaintiff in the lawsuit, former Google engineer Kelly Ellis, claims that despite having four years of experience before joining the company, she was placed at a tier of employment, Level 3, with recent college grads.* Meanwhile, soon after she was hired, a male engineer with the same amount of experience was placed at a higher level of employment, Level 4, making him immediately eligible for higher compensation and potentially more stock options, bonuses, and promotions, according to the lawsuit. Google admits this is a shortcoming of its current methodology and says it plans to dig deeper into the issue of leveling this year.

If the current analysis is inadequate, why share a counterintuitive statistic, drawn from a single job category, that appears to undermine the complaints of thousands of women who have worked at Google? The data point certainly traveled far. “Google Moves to Address Wage Equity, and Finds It’s Underpaying Many Men,” the New York Times titled its otherwise decent analysis. “Google reviewed pay equity and learned it was underpaying men,” read the Washington Post’s story. While these articles offer a nuanced picture of how bias can impact job placement, Google must be savvy enough at PR to know its highlighted stat would dominate headlines, even as it kept nearly every other detail of its findings private. Whatever Google wanted, the result was a raft of headlines that made it look like the company either doesn’t have a pay equity problem or, if it once did, it overcorrected for it and now has to correct again.

On top of the class-action lawsuit, Google is currently in the midst of a U.S. Department of Labor investigation into whether it underpaid women across its entire workforce. The company has resisted fully complying with the inquiry, arguing that forking over all the data the government requested would be too expensive for the company and too privacy-invasive for its employees. Coming from a company that had $136 billion in revenue last year and powers a lucrative ad-targeting business with its users’ data, both claims are a bit hard to swallow.

Google’s workforce is 69 percent male, and its tech workforce is 79 percent male. With such a large proportion of jobs occupied by men, it’s not surprising that the company found an unexpected wrinkle in a single category. In fact, it’s actually quite … intuitive. The company hasn’t released data about other job categories. Were those findings less surprising in Google’s eyes? After all, calling a statistic about men being undercompensated “counterintuitive” implies that the more intuitive statistic would be that women are the ones who are underpaid. It may be common for women to be paid less, sure, but setting that as an assumption risks further cementing the problem of wage inequity.

Thousands of employees at Google offices worldwide walked out last year in protest of how the company has handled sexual harassment and assault claims among its top male executives. Choosing to share a sliver of a report that appears to show that Google isn’t underpaying its female workforce isn’t the way a company that’s publicly grappling with discrimination issues should act. If this is indeed what Google found, why not wait to share the results until after a more complete analysis is finalized—or at least share the whole report and raw data?

Or maybe all that transparency would be just a bit too counterintuitive for Google to handle.

Correction, March 6, 2019: This article originally misspelled Kelly Ellis’ first name.