It's not very important, when you focus on outcome, whether or not a group of people conspired to create that outcome, or if that outcome was simply a consequence of other factors. The impact here is that straight white men over-represent the software industry as a whole, and there are things we can do to fix or improve that.
It's a super bad idea to hire anyone who is not qualified for a position, but I'm not going to pretend there's any kind of even remotely objective or precise way of determining who maximally fits into that position.
Instead, it seems optimal to acknowledge that there will always be more qualified candidates than there are open positions, and once you've found each qualified candidate, selecting the candidate that brings the largest difference in perspective (regardless of representation group) will be the best candidate. Given the saturation of straight white men in the SWE field, the odds that another straight white male will give the largest new perspective is not super high (though it is not zero).
The "action" here, if we need to walk away with one "thing" to do, is to saturate your pipeline with candidates from very diverse backgrounds, and then select the best candidate. It's a bullshit move to say, "only straight white men applied" if you did no work at all to reach out to other communities explicitly.
This is a wrong assumption, in my experience. As a rule of thumb, for not-principal/staff SWE roles, I would estimate that it takes:
1. 10+ resumes to find someone worth phone-screening
2. 10ish phone screens screens to find someone worth an in-person interview
3. 3ish in-persons to find someone worth an offer.
In other words, a hiring manager has to look at 300+ resumes to find one qualified candidate. So... imagine you get a reference from someone you trust. You go from 1 in 300 odds of finding someone who is basically qualified to 1 in 3.
This is after recruiters have pre-screened the resumes, btw. The candidate pool for Step #1 excludes all the people who apply to a senior SWE role with no Github portfolio, no relevant claimed skills, no degree and no work experience other than Burger King.
I'm curious to hear other people's experience, but I've literally never been in the position of "Do we hire candidate A or candidate B for this tech role?" It's always "Do we think A is good enough, or should we keep looking?"
> but I'm not going to pretend there's any kind of even remotely objective or precise way of determining who maximally fits into that position.
Sure, defining who is "optimal" is challenging but that's a cop-out. The situation is usually: Person A can't finish FizzBuzz (literally FizzBuzz) in 45 minutes in any language in coderpad, while Person B can do FizzBuzz, some easy recursion problem and maybe some kind of stats brain teaser. There is no world in which both of those candidates are "approximately the same".
Every single position I've ever hired for has had multiple people make it all the way through, often more like 3-5.
> Every single position I've ever hired for has had multiple people make it all the way through, often more like 3-5.
Do you hire for tech roles? Software engineers, data scientists, analysts and the like?
Also, having 3-5 people make it through your process isn't meaningful. You can always find n people make it all the way through if you wait long enough. The problem just that the (n-1)-th person will generally take another offer by the time you find the n-th person.
Like I said, I have never been in the position of having two qualified candidates at the same time for one role and having to pick which one I like better. My experience, by the way, seems to match that of everyone I'd discussed this with in real life.
You have to proactively attempt to break this cycle in order to make any real progress, and a few ways to do that include being intentional about how you find qualified candidates (growing your recruitment pipeline to source from non-traditional groups), and changing your interview process in ways that allow for candidates who are qualified to move forward together, rather than do some kind of false stack ranking of your candidates, creating this incorrect notion that "only one" person for any given role is qualified, which is objectively untrue.
Stack ranking is bullshit when applied to employee performance, why would it be anything but bullshit when applied to the interview process?
Your "action" doesn't fit with the reality I've experienced.
The world of ideas is so high-dimensional that sex & race impact very few of these dimensions. Most straight white men are not the same.
By what metric is software dominated by white men? Are we counting Indians as white?
But assuming 100% of white software developers are men, 54% is still not within the normal range of the word "dominated".
The power base in the US is white middle class. It's not just software that is dominated by white men. Law, medicine, construction, management at corporations. Not just white men. Privileged white men and the few "diverse" people mostly come from a level of privilege that would make my white lucky ass sick to my stomach.
Nice comment.