How Redfin Pays Women and Men

In America this year, women are earning 78% of what men earn. Some of this may be due to gender differences in income-seeking behavior, but much of it is due to sexism. It’s hard to advance the idea that capitalism is a force for good when it has so often been this unfair.
To narrow the gender gap, some employers are publishing data on employee pay. Redfin has now reached a size where we have a meaningful number of employees in similar roles, and so today we are releasing our own data on the relative pay of women and men at our company.
Our analysis looked at every level for which we had at least five employees in any given location. It found consistently equal pay across all roles. For public disclosure, we consolidated the engineering roles into broader categories like “all the developers and testers in San Francisco,” as a way to avoid highlighting the pay differences, however small, between a few individuals in a small group.

Because many Redfin employees don’t belong to a group of five or more peers in one location, Redfin employs more engineers and agents than we account for here. If there are only three mid-level software testers in San Francisco for example, we don’t have enough employees to make a fair comparison between men and women, and those three people are excluded from the comparison entirely.
Of note, the compensation numbers for agents do not include bonuses, which are set by the same plan for all agents, with the amount paid for each transaction based on the customer’s satisfaction; for software staff, we included target bonus amounts with the salary, because these target bonus amounts, like salaries, are set by human beings.
And while the data show that there are no significant differences in pay within a role, we’re still far from having an equal number of women and men in high-paying software engineering roles. Overall, 37% of Redfin’s engineering managers are female; 25% of all developers, test engineers, and operations engineers are female. Our first priority is hiring the best person for the job, but our chief technology officer, Bridget Frey, hopes to bring these numbers to 50% by 2020.
Releasing such data is easy to do once, after we’ve seen that disparities are limited, but doing this once creates an expectation that we’ll do it in the coming years, when new disparities between genders can always emerge.
Now of course some disparities will not reflect bias; some will reflect differences in employees’ value. It is entirely possible that next year three female entry-level software developers will be more capable than three male entry-level software developers, or vice-versa. Their pay will not be the same in that case, but it will be fair.
What is less likely is that, year after year, and in role after role, women in the same role will earn less than men in that role for reasons other than gender bias. We are thus promising to look for evidence of bias and address any problems, without promising to bring everyone in every role, every year, to exactly the same level of pay.
This is one way that society progresses, not by pretending that we don’t make mistakes, but by looking for mistakes carefully and addressing each one. And this is how businesses make more money, by figuring out which employees deliver the most value, especially the ones whom the rest of the market has overlooked, and then paying accordingly.
Beyond gender bias, there are other problems in how our society hires and pays people of different races, sexual orientation, and age. Though this analysis focused on the two most numerous, easily identifiable groups of employees, men and women, we want to be clear that we have a responsibility to be fair to everyone.
Many thanks to Redfin’s Women-in-Technology group who asked me if we could publish this data, and to Joscelyne Gray at Redfin for compiling it.
Silicon Valley often talks about disrupting this market or that, but women at Redfin and across the technology industry are now in a position to disrupt the oldest institution of them all, the ‘ol’-boys-network, using the weapons favored by disruptors over the past 20 years: data, transparency, the Internet, social networks, tools for anonymity and accountability, and a willingness to shake things up.
The result may be one of the best things technology ever does to capitalism.