Using Analytics to Close the Pay Gap

Using Analytics to Close the Pay Gap

The pay gap is the fact that women, on average, are paid about 80 cents for every dollar that a man is paid.  The pay is even less for ethnic minority women, with Hispanic women averaging about 54 cents for every dollar a non-Hispanic white male earns, and African American women earning about 63 cents per every dollar a white male earns.  With the recent allegation from the U.S. Department of Labor against Google claiming that they significantly underpay women compared to their male counterparts, Google has released their pay analysis methodology to show their strategy to close the gap.  Based on their analysis, Google refutes the claims.  Thanks to the rise in people analytics in the industry that I mentioned in my last post, Google is able to back up their claim with data and show they’re trying to close the gap.

So what is their way of closing the pay gap?  Well it focuses heavily on data analysis and leaves compensation review blind to gender.  Here are the ways that Google suggests structuring and checking for pay equity:

Step 1: Set a compensation philosophy

This is what companies want to reward.  At some it is pay for performance, or pay for tenure.  This step is deciding what pay should be based on and how the company wants to be positioned in the market.

Step 2: Structure your pay process

If you currently have no system in place this is where businesses should start. This is analyzing what jobs you have open, will have open in the future and what roles exist in your organization.  Then you decide what the fair market value of these roles are (how much would they get paid else where) and where your company should fall in this range.  This is comparing your pay to the market pay data.  During this step you should also group your roles to other comparable roles, for example an entry level analyst in marketing may be comparable to an entry level analyst in finance.  Then every time you decide on new pay (new higher or promotion) reference your pay structure. 

Step 3: Conduct a pay equity analysis

Conducting a pay equity analysis will help show whether or not you have factors influencing pay that you wouldn’t want to influence them, such as gender or race.  This step is where using analytics starts to show.  You will have to standardize your compensation variable and one way to do so is using pay ratios.  For example, Job A has a pay target of $100 and Person X in that job makes $90. The pay ratio is 90% (90/100). Job B has a pay target of $110 and Person Y makes $99 in that job. The pay ratio is again 90% (99/110). Using pay ratios, we see that Persons X and Y are paid similarly relative to their jobs, even though their actual salaries and pay targets differ.

Step 4: Identify variables to test

This is when you will control for certain variables to see if gender is correlated to your dependent variable/outcome of pay ratios.  You will control variables based on your compensation philosophy.  For example, you may want to control for things like seniority, performance scores etc.

Step 5: Analyze the data and look for variance

Now it’s time for the actual analysis, which will require someone who is highly skilled in statistics. Google provides a tool and detailed steps on how to do this along with examples.

Step 6: Act on your findings

Doing all of this analysis would be a bit pointless if you or the respective company didn’t act on it.  Implement changes, dive deeper into unconscious bias's that may be causing the discrepancies and consider pay increases. 

Using data and analytics to help create equal pay is absolutely an essential thing all companies should be doing.  A lot of this will negate other factors that continue the pay gap (i.e. the “what was your previous salary?” question).

Based on this detailed system of pay analysis, it will be interesting to find out how the U.S Department of Labor’s claims against Google stands up in court.