Measuring Diversity, Equity and Inclusion

If you’ve been in business long enough you know the phrase: if you can’t measure it, it doesn’t exist. Diversity, on the other hand, has been measured by companies for years, and it still doesn’t exist.

The following are current basic metrics of diversity:

1. Pay equity among diversity groups(sex, race, ethnicity, age, etc.)

2. Recruitment pipeline diversity: Percentage of diverse people among applicants compared to percentage of diverse people among final candidates and new hires.

3. Retention:Attrition level of each diversity group compared to the company’s general diversity level.

4. Advancement: Promotion rates of diverse people compared to those of the entire company vs. the rates of promotion within their own groups.

5. Representation: Percentage of people from each diversity group among job types(engineering, administrative, etc.) and job levels(junior level, senior management, etc.)

Here’s the problem with most of these metrics. They only measure the results, not how they happen. What you need is a set of metrics that helps predict outcomes as well as evaluate unconscious bias against minority groups. Better to analyze it sooner rather than later, before it impacts minority employees’ attrition rates.

Research(hyperlink) shows that feedback and training are the least effective ways of combating unconscious bias. People spend years in therapy trying to overcome beliefs that negatively impact their lives. Yet we’re trying to get people to change their beliefs overnight about something that doesn’t even impact their everyday life.

It’s like telling insecure friends that they should simply ‘be more confident.’ Doesn’t help much.

Instead, use strategies that don’t require people to catch themselves being biased. Focus on processes, not people or awareness. The right processes will nudge people to gradually change their biases, whether they are aware of them or not.

In Daniel Kahneman’s book, Thinking, Fast and Slow, he describes two systems of thinking. The first is the Automatic System, which “is or feels instinctive, and it does not involve what we usually associate with the word thinking.” Examples of the Automatic System at work include smiling when seeing a puppy, ducking when a ball is thrown at you, and subconscious bias and discrimination.

The Reflective System, on the other hand, is deliberate and self-conscious. People use it to decide which college to attend, where to go on vacation, and(under most circumstances) whether or not to get married.

The point is that it’s not easy for people to change their automatic systems; they need to deliberately choose to do so. If you try and force them to change their biases towards BIPOC, for example, they might just say they agree but not do anything about it.

You can however, nudge people in the right direction without taking away the individual’s freedom of choice. One example of a nudge is placing healthy foods in the front of the fridge, while placing junk food further in the back, making it less visible and harder to reach. Nothing is physically preventing individuals from eating what they want. The food is merely arranged in such a way that makes people less likely to choose unhealthy options. Nudge: Improving Decisions about Health, Wealth, and Happiness is a good reference if you want to learn more about creating nudge systems.

Creating this type of system, again, requires the right processes. Here are some metrics you can use to measure the process of diversity:

1. Amount of time recruiters and hiring managers spend interviewing candidates from diversity groups.

2. Amount of time diverse people have to improve their performance before getting dismissed.

3. Amount of money the company spends training people from diversity groups.

4. Hours people from diversity groups spend learning.

5. Percentage of people from diversity groups applying for promotions compared to the majority of the company.

6. Average salary increase across diversity groups compared to the majority.

These metrics are harder to implement than the basic ones, since some of them need more information than you typically record.

Let’s take a closer look at each one:

1. Many recruiters face a cursed cycle: they need to find people with specific experience for a non-diverse industry. For example, all recruiters looking for African American female engineers are talking to the same one percent of candidates. But they can’t expand their pool because they can’t agree on experience requirements. The problem isn’t a racial bias per se, but a bias against diverse experience.

To properly implement this metric, you would need to either record all interviews conducted or analyze them throughout an extended time period to notice a pattern. But if you implement it right, recruiters and hiring managers can keep their interview times in check to give equal opportunity to all candidates. Let recruiters and hiring managers know that they don’t need to be afraid if they inadvertently demonstrated bias during the selection process. That way they are more susceptible to feedback.

1. If you keep your disciplinary processes and PIPs on record, you can easily measure this data. Look into the dates of each PIP and warning, then compare the timing between dates. Documenting this information can come in handy later on, especially against lawsuits.

3 and 4: Both these metrics are rather easy to implement, as you can always pull records out of the financial department. From looking at the data, you may find that employees from diversity groups get less training because they tend to not ask for it. Make sure that everyone gets comparable training hours throughout the year, so that the training budget isn’t spent on the same group of people.

5. Unless you use your ATS to track internal hiring, which not many companies do, then you probably don’t measure this metric. To start, see if people from different diversity groups apply as often as those from the majority. This will show you how comfortable the company’s environment is for people who are less likely to raise a hand. Find out how to encourage these people. If they get rejected for a promotion, make sure they receive encouraging and actionable feedback.

6. Pretty much everyone is looking into pay equity. However, it’s not enough to realize that the gap exists. Start by tracking the average salary increase across all employee groups in percent and monetary value. Present the numbers to managers before they finalize the annual/semi-annual cycle, to make sure they recalibrate their data.

Emphasize that you’re looking for fairness, not preferential treatment. Do this by carefully examining monetary value. Let’s say you have two employees doing the same job, but one makes more than the other. They both perform equally well, so you decide to give them a three percent salary increase. While this seems fair from a glance, the person making more money actually earned the bigger raise, since it’s based on the original salary.

Providing managers with this information can give them a nudge in the right direction on giving employees equal treatment.

It’s impossible to change people’s habits overnight. Building proper processes and a system of nudges using these analytics can help you achieve a sustainable change in peoples’ biases and their behavior.