One of the key insights I think people are missing is that for most roles, not having a diverse slate of applicants in a country as diverse as America is a "process smell." The smell is that you're not casting a wide enough net to even know whether you're getting the best applicants.
Some easy examples: if you're hiring for most programming roles and you don't have any qualified women applying, your pipeline sucks.
If your company is in California, and you don't have any qualified Latinos applying for most of your jobs, your pipeline sucks.
Et cetera.
If you're not casting a wide enough net to find qualified Latinos on the west coast(!), I guarantee you're also missing qualified white men in whom you would be interested. Ultimately, this is why Microsoft cares; they have an interest in their overall process being the best it can be.
In 2021, 9% of computer science graduates were black (1). 18% were women (2). These numbers are the ceiling of what you might expect for a senior pipeline, since women attrit at a higher rate than men (3).
In a typical senior hiring pipeline for a Microsoft-scale company, while you'll have loads of applicants, you might have a dozen who make it through initial screening and are at least qualified on paper. I've been a hiring manager at companies slightly larger, and slightly smaller, than MS, and this was true at both.
So of those 12, you'd expect perhaps 1 to be black, and perhaps 2 to be women - optimistically. But every company, but especially high-profile fortune 100s, are trying to increase their D&I numbers. Qualified minority candidates rarely come through applications - instead, they're recruited, since everyone wants to somehow turn that 18% into 30% in order to get their D&I-linked bonus.
Blaming a hiring manager or company for their "pipeline" when they fail to hit impossible targets is absurd, and mathematically dishonest.
- as is frequently noted here, programming positions do not always require CS degrees.
- companies that produce software typically are composed of lots of roles. Microsoft has 100k engineers, but even more of their staff is not engineers. (That is: less than half the jobs at Microsoft are "engineers," and some of their engineers do not have CS degrees.) Many/most of those other roles translate quite well into tech. Just to take a few functional examples, the Finance function under the CFO, the HR function under the top people officer, the sales function, and the marketing function are not typically run by CS grads or staffed with programmers. Companies do not have to hit their DEI targets solely in their Engineering functions.
But back to your assertions, you seem to agree with me.
> So of those 12, you'd expect perhaps 1 to be black, and perhaps 2 to be women - optimistically.
Yes, if this is your slate for a programming role (again, assuming you are not somehow running a pipeline that somehow misses the enormous Latino population), you are probably doing okay. (Microsoft et. al. are able to aim for better than okay, but that's a business decision.)
Edit: I want to expand a little on the Latino/Hispanic component as well, as it's common in debates on this site to exclude them. But add them back in using the numbers on your first link. So we could expect 1 Black applicant, 1 Latino applicant, and 2 women (optimistically!). That's a third of your set of 12, using your data sources, for a programming role. This is optimistic, but nobody managing a F100 is asking their leadership to do easy things.
But it does sound like you generally agree that if you get sets of 12 applicants and they are all consistently white men in each set that there is a process defect? Similarly, if you are hiring for another role, your definition of a "diverse" set of applicants would be different.
> Qualified minority candidates rarely come through applications - instead, they're recruited
This is the activity that management is trying to encourage. Given industry history, it is not surprising that companies have to put in extra effort to try to change the perception of who is welcome. Microsoft is not new to this, having spent tons of money rehabilitating the image of Windows as an insecure, unreliable OS. Fixing a business process takes work, this is not surprising.
> But add them back in using the numbers on your first link. So we could expect 1 Black applicant, 1 Latino applicant, and 2 women (optimistically!). That's a third of your set of 12, using your data sources, for a programming role.
Let's please keep this conversation mathematically honest. I don't know if you're intentionally trying to misrepresent statistics, or if you don't understand them, but:
- for three non-mutually-exclusive groups (black, latino, female)
- given participation in a set (1, 1, 2)
- you cannot sum their independent participation (4) to yield "a third"
Beyond that, you appear to intentionally miss the key terms "ceiling" and "optimistically." Given a sample size of one pipeline (the OP), if you optimistically might get a single applicant that meets a given DI quota, it's very likely that you'll get zero.
The original article explicitly stated zero, and the numbers make it clear that zero is a very likely result.
That's the trouble with small sample sizes. I'm not sure what point you're trying to make by arguing that, in some other circumstance, maybe it could have been one... or even two!
When the set is of 12, that's 4 of 12, which is exactly one third, assuming independent participation (which obviously is what makes the number optimistic/ a ceiling).
> Given a sample size of one pipeline
Fair! If the OP only ever hires one person, you are correct. At Microsoft's scale, they are more like "the house" at a casino, and so their ratios should more closely approximate the population.
I am making the larger point that yes, in one instance, this hiring manager could have experienced a challenge building a diverse pipeline, but that this experience is not generalizable to Microsoft or the industry as a whole. And the related takeaway that if your pipeline is routinely not presenting you with the significant plurality represented by "diverse" candidates, then you have a solvable process problem.
> When the set is of 12, that's 4 of 12, which is exactly one third, assuming independent participation (which obviously is what makes the number optimistic/ a ceiling).
I think this misunderstanding of statistics may be at the core of our disagreement.
Yes, 4 is one-third of 12.
No, you cannot expect overlapping, non-mutually-exclusive sub-populations of 9% and 18% to equal 27% of a total population.
If you're interested in learning why, read (1), but the TL;DR is: given three people, A: black female, B: white male, C: white male, if you sum sub-populations as you're trying to do, you'll get (1 black + 1 female = 2) from a population of 3, and you'll extrapolate (incorrectly) that 2/3rds of this population meets your criteria. I hope that helps explain why your expectation of the gender & racial diversity of CS graduates is inaccurate.
> No, you cannot expect overlapping, non-mutually-exclusive sub-populations of 9% and 18% to equal 27% of a total population.
Obviously, this is true. I don't know why you continue to exclude the Latino population, but again Latino CS grads (to focus on that narrow slice of tech jobs) are also ~7%.
> I hope that helps explain why your expectation of the gender & racial diversity of CS graduates is inaccurate.
But (as you have pointed out) we do know more about the sub-populations that let us infer that the populations are less overlapping than we might prefer. If you're interested, there are published demographic data that go into more detail than I have time to do here. But yes, I understand the concept that a Black Latina fits into three categories.
> you'll extrapolate (incorrectly) that 2/3rds of this population meets your criteria.
The basic idea here is simple: at Microsoft scale, their pipelines broadly should look like the collegiate exit pipeline. If you think that < 30% of programmers are women and/or Black/Latino, I think that assumption is the problem.
No matter what assumptions you make around the numbers, my core point still holds. Which is that if your pipeline is routinely missing X% of qualified applicants, your sourcing is not good enough to know whether you're even seeing the best candidates. That is as true if X% is 15% or 30%.
Using an analogy: if your company isn't seeing any applicants from states representing 15% of the population (say: Texas and New York), that says more about your sourcing than the talent available.
Right, but here we're talking about Microsoft, one of the most valuable/profitable companies on the planet. Microsoft doesn't have to make do with a crappy pipeline. In fact, they have the resources to make it harder for the rest of us to have good pipelines.
And again -- I am talking about process smells. Women, Latinos, Black people, etc. are gainfully employed in roles where they ship software. (If you work in tech, you know the "CS Majors only" objection is a red herring given the breadth of roles at large tech firms.) If your process is unable to find them, that speaks to your process specifically (because someone else did find & hire them!).
Understanding why your process is bad is a good thing that every team should continually work towards.
1. Microsoft salaries are on the lower end compared to other FAANGs, this is well known fact.
2. Higher paying firms like goog fb are able to sweep the labor market and get the most of folks from tiny population of well qualified minority candidates. Making pipeline issues worse for everybody else.
3. At this point if you are minority and is qualified - you have no problem getting top tier job and top tier pay relatively easy, especially when compared to asian male population.
1) You are correct, but this does not speak to the need for Microsoft to have sub-optimal processes in any aspect of its business.
Separately, all of the big tech companies including Microsoft pay significantly more than typical employers in places where there are no big tech offices (this is most of the country).
2) This is just incorrect. Those firms cannot sweep the labor market because their US cultures have historically been so dependent on face-time. Microsoft is able to hire engineers in Atlanta in part because Facebook and Google have yet to put down significant engineering presences here, for example. Most medium-to-large US cities do not have significant FAANMG presence and so are also ripe for picking. This is why e.g. Target is able to get great talent in Minneapolis.
3) This is a claim that is difficult to evaluate given that engineering has had essentially full employment for going on 20 years.
1. process is not suboptimal, market conditions are different. Why would any well qualified minority go to msf for 120k in Atlanta if they can get 240 in Mountain view + relocation bonus? Plus most of qualified minority candidates are hired from top tier engineering programs are across the country. Local market is irrelevant at this point
2. You are wrong again, facetime is irrelevant, because of relocation. Very few will stay at MSFT in Atlanta for 100k if they can double pay + relo to CA/Seattle. Google does have office in atlanta, btw.
Your concept of supposedly "untapped talent in Minneapolis" is wrong, because in Minneapolis metro is about 3.5M people and Target can tap from talent pool of 3.5M, while places like Bay Area can tap from global talent pool (7B people) who self-selects and specifically migrates to places like Bay Area/NY/Seattle. This is three orders of magnitude difference in talent quantity and quality.
There is a reason Walmart established office in Bay Area, and doesn't want to limit itself to the great talent pool of Fayetville AR
> Why would any well qualified minority go to msf for 120k in Atlanta if they can get 240 in Mountain view + relocation bonus?
Because their family is in the Southeast, and because that $120k leads to a better overall living standard than $240k in Mountain View. Especially if one is raising/planning on raising children, proximity to family is a really important variable.
A process that systematically skips people who won't relocate to the West Coast is suboptimal in a world where we know remote work works.
> facetime is irrelevant, because of relocation. Very few will stay at MSFT in Atlanta for 100k if they can double pay + relo to CA/Seattle. Google does have office in atlanta, btw.
Reed Hastings and Tim Cook, among others, disagree with you. Google does have an Atlanta office, but most of FAANG does not have significant engineering presence there (Microsoft does). Notable that Google expects engineers in Atlanta to come to their office, even though their teams will likely be in other cities. This is the canonical definition of face-time.
And again, $100k in Atlanta or Dallas is roughly equivalent to $200k in the Bay or Seattle. Add in externalities like family and climate (everybody does not like the Bay climate!) and it's not as straightforward as you make it.
> because in Minneapolis metro is about 3.5M people and Target can tap from talent pool of 3.5M, while places like Bay Area can tap from global talent pool (7B people)
If you think everyone wants to move to the Bay, I probably can't convince you otherwise. But please consider that some people a) don't like the climate b) are married to people whose jobs are elsewhere c) prefer a lifestyle that includes homeownership.
Some easy examples: if you're hiring for most programming roles and you don't have any qualified women applying, your pipeline sucks.
If your company is in California, and you don't have any qualified Latinos applying for most of your jobs, your pipeline sucks.
Et cetera.
If you're not casting a wide enough net to find qualified Latinos on the west coast(!), I guarantee you're also missing qualified white men in whom you would be interested. Ultimately, this is why Microsoft cares; they have an interest in their overall process being the best it can be.