[anyonecancode]

The way I think of this topic is that diversity should be a metric, not a goal. If the overwhelming majority of the people I'm interviewing, hiring, and working with look like me, then I have to ask, are these really the most qualified people? I mean, there's one of two possibilities -- either in fact there's something about people who are very similar to me that makes them the most qualified OR there's a lot of talent being overlooked and missed and my company is missing out here. Which one is more likely?

That re-framing is the easy part, of course. Actually making changes is hard. On one side we all know that if you make something a metric you run the danger of making it a goal, but on the other side the biases that lead to a lack of diversity are so deep and ingrained they can be very difficult to counter.

But hopefully this re-framing is at least helpful in explaining why a lack of diversity is an issue that matters.

[superbcarrot]

> I mean, there's one of two possibilities -- either in fact there's something about people who are very similar to me that makes them the most qualified OR there's a lot of talent being overlooked and missed and my company is missing out here.

The third and more likely option is that you're sampling from a population that looks mostly like you. If software developers are 90% male and men and women are equally competent, you should still expect to hire 9 times more men than women even if you could pick the best candidates every time.

[anyonecancode]

But where did you get this measure of the population from? I'm skeptical that "population of software engineers I'm familiar with" is equal to "total universe of software engineers." And of course the "equally competent" -- well, if we're already deciding ahead of time that we know who's competent and who's not, I think that might be part of the problem, no?

[myWindoonn]

To reply to both you and the original question, I talked about Goodhart's Law [0] recently with upper management on the topic of D&I. Since metrics become useless as they are used to define goals, there is a constant need to rise above expectations. To paraphrase one executive, D&I is a process and not just a checkbox.

[0] https://en.wikipedia.org/wiki/Goodhart%27s_law

[threatofrain]

Why should it be a metric if it's not part of our goals? The problem is there is no language for negotiating the relative weight of values.

For example, how should business success weigh against the social justice problem of black access to technological opportunity?