Diversity Debt in Leaky Pipelines
We Analyzed Millions of Recruiting Data Points. This is What We Found.
At Atipica, we are advancing the science of hiring by finding the right balance of people knowledge while utilizing the power of machine learning. It is this combination of human and artificial intelligence that makes us unique.
While we focus primarily on the quantitative power of talent data, we simultaneously work with our business partners and clients to help them gain the insights needed to improve their qualitative goals — recruiting, hiring and diversity.
From startups with fewer than fifty employees to mature tech post-IPO clients, we found similar trends across the industry. But most importantly, we found that each company had different challenges unique to their own history and growth stage.
Our suite of products focus on the talent lifecycle and the power of data.
But given that our founder and CEO is a woman with a career in tech since the age of seventeen and a founding team member of Project Include, diversity is part of our DNA.
We care about this. We have created technology to help diversity and inclusion.
This is why we are introducing a series of blog posts to look at how data-driven approaches to talent acquisition can help diversity and inclusion while also providing industry benchmarks on how your tech peers are doing.
Note: The algorithms to determine race and gender are proprietary to Atipica. We have gathered public, location and private data to feed our gender and race models in talent acquisition. Furthermore, by employing machine learning, our accuracy rates are upwards of 96%.
Across all roles:
Males are overrepresented in the applicant pool by an average of 67%.
However, among our clients, women represent an average of 51% of the hired pool.
The trends we have seen is that hired pool does end up being with balanced gender representation across companies. This is an example:
First things first. Across all companies, regardless of size, companies are doing a better job of attracting and hiring female talent now than 5+ years ago.
Let’s take a look at another case study, Company Y.
Amazing! Tech companies are making progress, right?
To a certain extent, yes.
But these changes in hiring trends are are due to the combination of company inputs and organizational biases among different teams, as well as significant workforce behaviors along gender lines.
Here is what we found:
Men have the power of option— male applicants reject all companies at a significant higher rate than female applicants.
Women hit the hiring glass door — they are rejected at above the rejection rates of the company.
But, because women accept roles at a higher rate (2.5x) than men, this results in a more gender-equal hired pool.
This is true for the majority of tech companies — regardless of roles.
The gender diversification of tech employees is mostly due to…. female applicants. They are more willing to accept roles if given an offer letter.
General analytics can provide interesting insights as seen above. But each client is tells their own story through their structured data found in their Applicant Tracking Systems (ATS) and unstructured data (resumes, cover letters, etc).
Our proprietary models for gender and race are focused on this — the talent lifecycle. We can pinpoint exactly where in the funnel, what role, whom, what location and other hiring factors contribute to differences in the leaky pipeline. As well as how you measure against other players in the industry.
We are able to do this automatically, in real-time, without self-reporting surveys from your candidates and interview panels.
Taking all of this into account, Atipica is able to explore the biases (and implied -isms) that lead to higher rejection rates for women. By using our suite of products, we help companies restructure recruiting operations and processes to improve their talent goals and employ innovative solutions to their hiring goals.
Our next steps at Atipica are to take a look at whether the difference we find are by chance or if they are statistically significant, and if so, how does the impact of diversity throughout the whole talent lifecycle — from recruiting to employee engagement to culture surveys to general company data — affect the end goals. We look to employ a similar approach to this research:
“A difference in performance evaluations favoring men but accounting for only 1% of the variability in scores can produce an organizational structure in which 65% of the highest level positions are filled by men.
Thus, relatively small sex bias effects in performance ratings led to substantially lower promotion rates for women, resulting in proportionately fewer women than men at the top levels of the organization.” (Martell et. al)
Care to learn more?