Universal Categories: A Proposal
A continuing pain-point in Diversity reporting, is a consistent taxonomy for categorizing individuals across all types of businesses within a specific sector (eg: Tech). Why does this matter? Most obviously, because it’s butkis for an organization to present itself as having gender parity in its workforce, when +90% the product creators are men and +90% the customer service and admin roles are women.
We love data because it gives us objective dots along imaginary lines we can then draw, to tell stories with. Stories derived from the objective data are what matter the most—not the data alone and devoid of context.
As a starting-point for the storytelling from Tech’s diversity data, we’ve thus far evolved to [Tech & Non-Tech] as our horizontal categories set. Do those categories tell the most complete story? Nope. [Engineering & Non-Engineering]? Nope. Yes: the disparity of PoC and Women in Computer Science is an issue, and those two bucket-types are a starting-point for that unique story. It is just one issue, though. How diversity within any sector impacts the world around it is a much more complicated equation than any bucketing of roles, though—and taking our diversity game to the next level with a meaningful framework of categorization for context, and past “OMG, they released their numbers—yay!” it’s high time to do.
Since my purpose for EqualTogether was to collect data to tell those stories, I created the below as my own jab at evolving that contextual framework—and today, I’m throwing it out into the world to share, far and wide. My hope is for organizations consider adopting a similar, flexible, and sector-consistent schema of categorizing the roles each individual plays in supporting the company they work for. Published data only matters when it can be consistently compared against other parallel data sets—and for those parallels, categories & hierarchies must be consistently defined & adhered to.
Data Tells A Story—My Proposed Taxonomy, Explained
A great example of where universal categories could have helped the public understand a Diversity & Inclusion report, was Apple’s 2014 first-ever diversity numbers press release. Apple shocked everyone with an on-paper industry lead in the percentage of PoC ICs (Individual Contributors) on its workforce. The Tech sector is currently trending at an average of ~2% Black ICs and less than ~1% Hispanic ICs, yet Apple blew its brethren out of the water in reporting 11% of its 93,000-person workforce as Hispanic and 7% of its workforce as Black. WOW!! But wait, that report failed to offer vertical or horizontal distillations of those numbers… so in effect that 5x best-representation of Black workers doesn’t really mean anything.
Apple manufactures hardware, makes software, does media sales (all content on iTunes and Apple Music, plus all 3rd party apps), and has a sprawling retail enterprise. Google (erm, are we calling them Alphabet yet)? Software mostly, and then all their esoteric side/spinoff things. Twitter? Software. Salesforce? Software (sorry, Marc!). Instacart? Some software, but also a massive fulfillment team for service execution. ZipCar? Software, plus unique customer care and fleet management offices for each metro area they serve—so again, service fulfillment. Postmates? Instacart? Amazon? Good lord, where to start. You get it.
So with Apple, my speculation is that yes, they have lots of Black workers in their retail stores and Hispanic workers in their factories… but the usual whitewashing at HQ, for Product Development, Sales, and Ops. It’d be nice if that could be more obvious, though—or my assumptions challenged with a far more interesting truth that could be pattern-matched across similar companies. To solve problems, we gotta have consistent data stories as starting-points. Universal taxonomies & data collection standards, are requisites to achieve that consistency.
How Its Done Today
EEO diversity reporting has categorized ICs in the most generic fashion possible, that maps to every American workplace possible and across all business sectors. The highest level of filtering businesses themselves is Farm and Non-Farm, and it only gets more irrelevant with individual contributor categorizations. Professionals, Technicians, Sales Workers, Craft Workers, and Operatives are my favorite obtuse IC categories.
Tracy Chou made early waves in the Tech diversity space, when she created her adhoc/shared spreadsheet to quantify gender disparities at her employer, Pinterest. If Pinterest were an anomaly I suspect she would have just quit, but they’re not—and Tracy wanted to tangibly quantify what most of us had already (quietly) pattern-match established as the norm in workplaces everywhere—Tech, Finance, Airlines, Energy, etc—that there are chick roles and dude roles, but for no logical reasons. Want to fix a problem vs endlessly kvetching about it? Great, start with data. Tell the story. Objective data is the yellow brick road to the Kansas of solutions. The Emerald City is a great idea, but Kansas is where we really need to be. :)
[Engineering & Non-Engineering] was the set Tracy used to illustrate the epic disparity of women in “making the thing” roles at Pinterest—which was what made the most amount of sense for the kind of company Pinterest is, and with the types of roles for ICs it subsequently offers. Tracy is also an engineer, and was rightfully annoyed by what she observed through her own experiences, to be clear gender disparities on most engineering teams across the entire Tech sector.
The categories set eventually evolved to [Tech & Non-Tech], yet that still remained too broad to say much while also diminishing non-Tech roles, and overlooking the point of why diversity matters so much: it impacts how the products themselves are shaped, and the distribution channels they fail or flourish through.
Employee happiness and workplace culture are critically important, but how products are shaped and distributed is how our 9–5 doings shape the world outside our bubbles of sector-specific workplace culture—and that matters.
A Clear, Holistic Tech Sector Taxonomy
As a non-engineering-though-product-shaping IC in Tech, I found myself and others asking the following questions: Did Tech include UX and Product Marketing? If not, is that because Product Marketing really isn’t that important in a product’s reach or in how it’s shaped, or is it just a personal bias against marketing roles as necessary or important (a common bias among more technical folks)? If a company’s Board of Directors wields the epic influence over its overall direction that they do (on an aside from being a status symbol to both its individuals and a representation of corporate values), how come the BoD’s individual directors aren’t included in the Management category, too—and what does Management mean… just the real execs, or all the middle-managers, too?
With businesses like Instacart or Amazon or Apple, how do their fulfillment employees fit into the categories? Should they even matter? If not, why?
EqualTogether + Storytelling Inclusion As Statistics
The most obvious schema for me to adopt with EqualTogether were Radford’s categories. While many organizations do use the Radford category sets to inform pay decisions, it’s uncertain who does and who does not. Beyond that, it’s also uncertain if Radford’s taxonomy is the most user-centric for use in comparing data between companies, for diversity (and not pay) evaluation—or if they’re even relevant to the types of businesses evolving in the fast-moving Tech sector. My guess is, no.
US Census or Bureau of Labor Statistics? It’s on the US Congress to update the Census categories (in the most epic design-by-committee fail, I’ve yet unearthed), and that only happens every 10 years. The US Department of Labor Statistics, I wrote about in this article after it blew my mind with its oversights, relics, and “design by government” complexity.
Because categorization-updates impact how statistics are calculated and progress demonstrated between category change periods, and Radford’s goals are to demonstrate Q2Q salary trends, I also doubt Radford‘s taxonomy is structured to scale and grow to maintain story-consistency as new categories come and old ones go. Or, that Radford does any updates with much more frequency than the US Census Office.
Anyhow: so, this is a downloadable/printable PDF of the categories with their drill-down paths, that‘s pictured above. It’s been Creative Commons licensed, and fewer things would make me happier than to see it put to good use either as a conversation catalyst that leads to results. Please circulate it far and wide, discuss among your peers & internal teams, and let’s get busy!
