The Pipeline Problem
Another year has passed, and despite major commitments from various firms and monumental efforts by individuals, diversity in tech continues to see a sparsity of meaningful improvement.
A common explanation is the pipeline problem. It describes the challenge of hiring members of underrepresented groups for technical roles (women, people of color, veterans, LGBTQ), given that members of those groups are often less likely to pursue technical careers.
While this challenge has deep-rooted societal bases and a fair degree of legitimacy, it cannot explain the full gaps we see in tech. There are problems across the candidate life-cycle, which can easily become overwhelming for those trying to make a difference.
Emergence of Distributed Workforces
Further complicating matters is that technical workforces are becoming increasingly distributed across geography. More and more tech companies are building offices outside of Silicon Valley, and even more are offering co-working spaces and work from home arrangements. If looking across the candidate lifecycle for a company in general is overwhelming, looking at it across geography is just plain dizzying.
Much of this complication stems from the fact that different geographies (cities, states, countries) meaningfully vary in:
- the population base rates of different groups
- the labor force participation rates of different groups
- the technical working population of different groups
For example, Texas has a lot more Hispanic people than Washington, Minnesota has a higher labor force participation rate of women than Louisiana, and California has a higher rate of tech workers than Alabama.
All of the factors mentioned above make it difficult to determine how well a company is mirroring the technical workforce composition of where they work. At the same time, however, this variation also provides immense opportunity to drive improvement.
Imagine a case where a company finds that some of their locations have large populations of technical women that are not being fully tapped. Instead of dividing their resources for outreach recruitment across all their locations, with this knowledge they could focus solely on locations with the best expected results.
The pipeline opportunity, then, describes how varying sizes of technical workforces across different geographies can allow companies allocate their limited resources more efficiently.
The very same logic can be applied across the candidate lifecycle as well, where differing application and conversion rates across geographies can help companies pinpoint where their recruitment and selections processes need improvement.
So even though geographic variation of workforces makes everything more complex, it can also be a powerful tool to help improve the state of diversity in tech.
Applying this Approach in Practice
To apply this approach in practice to any organization that is geographically distributed, one can go through the following 5-step process:
- Identify which locations have much smaller percentages of employees from underrepresented groups than local workforce compositions would imply
- Of those locations, find which ones have application rates of underrepresented group members that are lower than the local workforce composition would imply
- Focus outreach recruitment resources on the locations identified in this way
- Every 6 months, revisit (1) and (2) to see if improvement has been made for any locations.
- Follow-up actions
- For locations with little to no improvement, change your outreach recruitment strategy
- For locations that improve to a degree such that (1) and (2) are no longer true, slowly reallocate resources away from that location to those that need it more
- For locations where (2) is no longer true but (1) remains true, review your selection process and quality of new applications received to dive deeper
This process can also be tailored to specific types of workers and departments. If a company wants to improve the representation of employees in technical roles specifically, for example, they can use numbers for employees, local workforces, and application rates that are specifically technical. These analyses are limited by the numbers an organization can access, but otherwise the possibilities are endless!
Closing Thoughts & Resources
Diversity in tech is a huge, multifaceted problem and I do not believe that any one company can solve it. I do believe, however, that in adopting more localized strategies where areas of challenge and opportunity are identified, companies can make smarter investments and slowly chip away at the disparities we see today.
Despite what I have outlined, I recognize that finding relevant datasets, and conducting relevant analyses is a large barrier to seizing this opportunity. To address this barrier, in the spirit of open source I have created some technical resources (R scripts and datasets) that make it easier for analysts to conduct the analyses outlined in this piece. This is a work in progress, and I fully invite others to use these resources, collaborate on the project, and start new ones of their own.In closing, I want to acknowledge that the pipeline opportunity I have outlined addresses only one small part of the problem of diversity in tech. Differences in treatment, upward mobility, and retention across groups all contribute to the disparities we continue to see as well.
Tackling these problems alone is not just overwhelming, it is simply impossible. With this in mind, I am truly glad there is a community of people who are passionate and working on these problems, and that the information age allows us to share our ideas and build a better future together. Slowly but surely, let’s keep trying to do better.
Disclaimer: This piece highlights my own personal work and thinking on diversity across geography, and does not necessarily reflect GitHub’s philosophy or projects on this topic.