もっと詳しく

In the tech industry, diversity is a hot topic. We’ve seen companies such as Facebook and Google commit to diversity, equity and inclusion initiatives in hopes of improving their company culture. But analysis shows that there has been no appreciable change in diversity for these allegedly DEI-driven companies over the past 10 years despite these efforts.

In the 2010s, many tech companies made their annual diversity reports public to demonstrate transparency. Still, a problem emerges when reviewing the data: While the reports often describe incremental gains or losses in diversity numbers, they often fail to analyze data that can help the company make tangible change.

Using data to dig deeper into systems that impact diversity

The data in these reports provide us with a statement of the outcomes. We can see if the company has had growth or decline in the measured areas of diversity — typically demographic representation. What we can’t see is why these outcomes are happening and at what point in the system we’re missing the mark.

Therefore, these reports miss out on the powerful accountability effects of measuring improved processes and systems.

To illustrate an example, we’ll look at a company’s marketing data.

When a large corporation is trying to optimize its marketing funnel, it can see data at every level of the marketing process. For example, they can estimate the number of impressions they receive from external marketing efforts, quantify the number of conversions from ads to their website and calculate how many website visitors convert to customers. This information is used regularly to optimize performance and increase sales — and it’s prioritized because leaders know that revenue generation is essential for the company to succeed.

Leaders can accomplish this quality of analysis at the hiring level by recording and analyzing data at every level of the recruiting system. Data sets would include demographic information for every point of interaction, including but not limited to:

  • The number of impressions received for a job ad.
  • Candidates identified through sourcing channels.
  • Candidates being presented for interviews.
  • Interviewer assessment results.
  • Job offers and job offers accepted.

Leaders could then ensure demographic representation at each level of the process is at least equivalent to the demographics of the available talent pool. The company would then need to prioritize using this data to make changes to the system regularly (not annually) to improve the results.

Recruiting is only one example, and there are other areas of business that can benefit from more informative diversity data.

Making information actionable

As a leader, you have a choice to make your data actionable by publishing information that can result in accountability and greater equity and inclusion.

Consider sharing each of the following:

  • Pay transparency and pay equity by demographic.
  • Engagement and inclusion data by demographic.
  • Rates of promotion by demographic.
  • Rates of retention by demographic.

Now, your general counsel’s skin may crawl at these recommendations. However, organizations innovating on talent, culture and equity are willing to wander into these territories of transparency and accountability.

Moreover, we have seen organizations that are willing to stand on their integrity wholeheartedly by publishing these data make significant progress toward building cultures of belonging that center on diversity.

Data transparency doesn’t equal data accountability

We often believe that if we can measure something, we can change it; however, measurement alone does not drive change. Therefore, it’s crucial to measure the right data elements while holding stakeholders accountable for measurable change.

Let’s consider how this is done in sales.

Sales teams measure individual performance based on their contribution to achieving revenue targets. They are required to get results or risk losing their job because failure can cause harm to the business.

If we did the same for DEI, we’d see leaders held accountable for getting tangible results quarterly or annually as part of their performance appraisal goals. But, unfortunately, putting out a report stating what occurred won’t inspire leaders to do more to change the processes that impact outcomes meaningfully.

Benchmarking against already low numbers won’t move the needle

Diversity reporting often benchmarks annual data against the company’s previous metrics, the industry as a whole, companies of a similar size or an entire geography (i.e., the U.S.). This form of measuring progress makes incremental progress seem like a bigger achievement than it is. Leaders can point to this data and say they are meeting industry standards, but if the industry’s rate of progress is negligible, it simply removes accountability for achieving better results. Organizations have not been getting diversity right for decades, why benchmark to those underperforming?

Let’s say this simply: Benchmarking to poor performance is a poor practice.

If you do a benchmark, at least seek to make a real improvement by benchmarking against companies performing in the top quartile when it comes to diversity representation. But, again, this sets the bar higher and challenges leaders to be more strategic.

More importantly, companies should benchmark against the available talent pool (by using Bureau of Labor Statistics data in the U.S., or by graduation rates by demographic and field of study) to determine underrepresentation within your geography, industry and job field.

For example, the rate of graduation among women computer science majors is dramatically higher than the representation of women in junior-level software engineering roles in most companies.

And yet, even BLS data can be flawed because it relies on several assumptions about who is eligible for employment and excludes whole swaths of truly employable people. So, you may also combine these benchmarks with local population data as yet another data set to show how effective you’ve been.

If tech companies publish data that not only describes what is currently happening in the organization but also gives context to how systemic bias impacts diverse talent who cannot access their organizations, these reports might motivate behavioral change and be more valuable for readers from other industries aspiring to improve their own organizational outcomes.