Many of us (s­peak­ing from the tech sec­tor where I work) think the sector’s work­place di­ver­si­ty isn’t very good. Specif­i­cal­ly, there aren’t enough wom­en. Large com­pa­nies  —  all the ones I’ve worked for, any­how  —  have goal­s, and gen­er­al­ly work hard at meet­ing them. Many com­pa­nies now say they care about di­ver­si­ty, and have goals around im­prov­ing it. But im­prove­ment is painful­ly slow; why? Maybe part of it is that those aren’t the same kind of “goals”.

How busi­ness goals work · When I say “large com­pa­nies have goals”, I mean that in a very spe­cif­ic way. Each plan­ning cy­cle, com­pa­ny groups and their man­agers take on a set of ex­plic­it­ly written-down goals for that plan­ning cy­cle. Goals are tracked in a sim­ple database and at the end of the year, each group/­man­ag­er gets a pass/­fail on each. The way that goals are de­fined and re­fined and agreed to and record­ed and struc­tured dif­fers from place to place; at Google and sev­er­al oth­er big high-techs, they’re called OKRs.

The per­cent­age of goal com­ple­tion that’s re­gard­ed as “good” al­so varies, but it’s nev­er 100%. The idea is that your reach should ex­ceed your grasp, and if you score 100 you might have been sand­bag­ging, choos­ing in­suf­fi­cient­ly am­bi­tious goals to make your­self look good.

Goal com­ple­tion is dead­ly se­ri­ous busi­ness among most man­age­ment types I’ve known, and the num­ber has a re­al ef­fect on ca­reer tra­jec­to­ry and thus com­pen­sa­tion. I don’t think it’s con­tro­ver­sial to say that in busi­ness, those things mat­ter a whole lot.

Goals are sort­ed in­to “output goals” (ex­am­ple: $100M in sales for a pro­duc­t) and “input goals” (ex­am­ple: five cus­tomer vis­its per week by ev­ery sales­per­son­). They can be tech­ni­cal too, around things like up­ti­me, la­ten­cy, and trou­ble tick­et­s.

In­put and out­put are not mu­tu­al­ly ex­clu­sive. In­put goals are at some lev­el more “reasonable” be­cause they are things that an or­ga­ni­za­tion con­trols di­rect­ly. Out­put goals are more ag­gres­sive, but al­so lib­er­at­ing be­cause they turn teams loose to fig­ure out what the best path is to get­ting that sales num­ber or up­time or what­ev­er.

Gen­er­al­ly, I like this man­age­ment prac­tice: Set­ting goals and mea­sur­ing per­for­mance against them. It drives clar­i­ty about what you’re try­ing to achieve and how well you’re do­ing.

Diver­si­ty goal ques­tions · Here’s a ques­tion: For any giv­en com­pa­ny, do its di­ver­si­ty goals work like reg­u­lar com­pa­ny goal­s? That is to say, do they go in­to the per­cent­age com­ple­tion num­ber? The num­ber that man­agers get judged on and re­ward­ed for meet­ing?

I ac­tu­al­ly don’t know what the an­swers would be for most high-techs, but I sus­pect it’s “Not of­ten enough.” I sus­pect that be­cause the di­ver­si­ty num­bers across the high-tech land­scape are uni­ver­sal­ly pret­ty bad, and be­cause the peo­ple in man­age­ment are gen­er­al­ly, you know, pret­ty smart, and will come up with re­mark­ably clever ways to meet the goals they’re get­ting judged on.

I’ve al­so ob­served that while the num­bers are un­sat­is­fy­ing in the large, there are teams who con­sis­tent­ly man­age to do bet­ter than oth­ers at hir­ing and re­tain­ing wom­en. And by the way, anec­do­tal­ly, those are good teams (with good man­ager­s); the kind who get things done and have low at­tri­tion rates and hap­py cus­tomer­s.

Here’s an­oth­er ques­tion: For di­ver­si­ty, should we be talk­ing in­put or out­put goal­s? I say: Why not both? I’m not ex­pert on the state of the art in build­ing di­ver­si­ty, but wher­ev­er we know what the equiv­a­lent of “five cus­tomer vis­its per week” is, let’s sign teams up for a few of those. And yeah, out­put goal­s. Let’s ask man­agers to dou­ble the pro­por­tion of wom­en en­gi­neer­s, mea­sure whether they do it or not, and leave the de­tails to them. The good ones will fig­ure out a way to get there.

It’s like this: If you claim you have di­ver­si­ty goal­s, but your managers’ ca­reers don’t de­pend on their per­for­mance against those goal­s, you don’t re­al­ly.



Contributions

Comment feed for ongoing:Comments feed

From: Nik P (Aug 18 2018, at 02:30)

I feel like it's also important to preempt the constant backlash of these discussions by always referencing diversity goals with an explicit postscript of "Without changing the hiring technical bar".

This is unambiguous from the inside looking out, but I've gotten feedback that it's not clear from the outside looking in. I hear it all the time from non-allies: "If you succeed in increasing diversity, you will make non-diversity employees wonder if the new diverse employees around them are actually not as good, and were hired for non-technical reasons. And that will harm everyone, especially those women/minorities who already managed to get hired before any diversity initiatives."

The thing is, I think it's a legitimate concern. Companies like Google and Amazon have VERY difficult and structured hiring processes, aimed at making decisions as data-driven as possible. And we would not and could not compromise them.

But that's not the case in the rest of our industry. Plenty of companies still hire based on nepotism or culture fit (where "culture" is liking IPAs and DOTA). Yet they are just as actively engaged and interested in "diversity" because for a lot of companies it's not about fixing what they see as a problem, or trying to improve their productivity, or diversifying their thought; it's only about virtue-signaling to the community that they care. But they are cargo-culting, and don't really understand why they SHOULD care.

[link]

From: Marc (Aug 18 2018, at 21:17)

I'm African American, and have been working as a software developer and manager since the early 80s in the SF Bay Area, currently working as a researcher at one of the major universities here. My wife also works as a software developer at the other major university, and has worked for PC software companies in past years, including the one where we met. Perhaps more interestingly, our daughter (who is what they call "mixed-race" these days) is entering her senior year majoring in CS at the university where I now work. We're hardly a "typical" family in this regard, but the thing about diversity is that it's never "typical".

I'm not very impressed with this concept of "diversity goals", they don't solve the problem, only quantify it. I am also amused by Nik P's comment about "VERY difficult and structured hiring processes, aimed at making decisions as data-driven as possible" used by Google and Amazon, yet if you look at their diversity numbers, they are actually worse than many other companies. The technical aspects of these processes were developed by the existing software managers to recruit new programmers out of college, most of whom, oddly enough, end up looking a lot like younger versions of the managers who came up with the process.

The thing is, the solution has always been obvious, I question whether any of these companies really want it solved. But, for example, when I first arrived in the Bay Area, there were exactly three companies that were considered "minority"-friendly (minority then including Asians, who were then almost as invisible as Blacks), HP, Xerox, and Silicon Graphics. HP and Xerox had corporate mandates for equal opportunity, and took them quite seriously, plus they both hired extensively from the military. Silicon Graphics was an outlier, having worked at both SGI and Sun, the difference was striking. Walk in the front door, the receptionist was often a Black woman, walking through the halls you'd see Black faces all of the time. The reason was simply that the CTO was Black, and well, there wasn't any risk associated with hiring Blacks and other minorities for technical positions. Sun, uh, when I was there in the late 80s, there were exactly 4 Black technical employees in a company of over 10,000. Nobody, was openly hostile, but, there wasn't much encouragement either.

Companies now like to talk about "diversity", they do a great job of hiring consultants, running training sessions, etc. They hire a few women and "underrepresented minorities" each year, have a hard time holding on to them, and assume the problem is that there just aren't enough of "them" out there, or that the ones they find aren't skilled enough. The real problem is cultural, the managers in these companies look for candidates with familiar backgrounds, experiences, and problem solving approaches. Many are confused by and automatically dismissive of those who do not fit their model. The solution is to change the culture, not impose hiring quotas on managers. The way to change the culture is to hire executives and managers who come from non-traditional backgrounds, not hiring a few gifted non-traditional recent college grads from elite colleges.

[link]

From: Rob Bray (Aug 19 2018, at 12:36)

What Marc said.

For a number of years, I made a fair amount of beer money on the side as an Employment Equity consultant, and as an "Outcomes & Logic Model" consultant too. (Outcomes and Logic models is what Tim is referring to in terms of outputs, inputs, etc.. It is a VERY highly developed art in the human services area.)

But Employment Equity consulting, well. I eventually had to quit, because the work, while insanely lucrative, was essentially the organizational equivalent of bottom tier prostitution, and it became impossible to live with myself any more (oceans of beer notwithstanding). Essentially, nobody actually wanted to change anything, they just wanted to virtue signal.

The problem, in the end, is essentially that of the Peter Principle. I have yet to meet a manager that did not firmly believe that he (usually he of course) had above average skills in recruiting and hiring talent (the Lake Woebegon effect). But in actual fact, nothing about being a superior technical worker indicates that you will have any capacity for leadership, staff selection and motivation, or any kind of strategic thinking. If you do, it is just random chance. (And as Dr Peter pointed out, if you luckily do, you will eventually get promoted into a job you actually do not have any aptitude for.)

How this usually cashes out, is that our technically skilled but likely management skills challenged executive, confronted with the task of hiring talent, concludes in effect that 'I am a superior coder/carpenter/accountant/salesman/ditch-digger/nurse, as evidenced by my promotion to this exalted position, therefore if I hire someone like me, I am ipso facto hiring a superior candidate.' It seems like a plausible quick and dirty hack, and we all know where you end up if you base your entire system on plausible-sounding quick and dirty hacks.

It would be great if there actually data-driven testing protocols that actually work, but there simply aren't. You can use all the Stanford-Binets, Myers-Briggs, Google/Amazon/Walmart etc testing you want, but the simple fact is that the whole field of psychometrics is, and always has been, a complete and utter bust at best, and a cruel pernicious racist/sexist/classist fraud at worst (cf eugenics, history of), with only the vaguest correlations with real world outcomes.

Now logic model, outcomes driven evaluation systems have their place, when correctly applied (I am something of an evangelist), but I am sorry, the only real world, proven (time and again) method of achieving organizational diversity is to make the biases work for you-- as Marc says, you gotta put some female/African/First Nations/LGBTQetc people in positions to make hiring decisions. So they too will hire "people like me." Not in virtue signalling positions in HR or PR, but in real actual responsible line management positions, making final hiring decisions they are going to have to personally live with.

And not worry so much about finding super skilled "qualified" ones-- the people most organizations are already promoting into management positions actually probably don't have a ton of management capacity anyway; it literally would be hard to do worse. (This is kinda how the military does things-- being a great NCO does NOT guarantee entry into officer school, even though there are a lot of terrible junior officers. The military at least recognizes its a whole different skill set.)

[link]

From: len (Aug 25 2018, at 09:32)

Diversity as a goal in conflict with profits as a must is too obvious to mention but it is as always the elephant in the canary cage.

Virtue signalling isn't. More often than not, virtue signalling is much like being a serving officer of her Majesty's Royal Guard: a sound practice for officers but a career-ender for staff and below. The word applied there is "weasel".

Success is all about your mates. See Major John Andre.

[link]

author · Dad · software · colophon · rights
picture of the day
August 15, 2018
· The World (112 fragments)
· · Politics (155 more)
· Business (109 more)

By .

I am an employee
of Amazon.com, but
the opinions expressed here
are my own, and no other party
necessarily agrees with them.

A full disclosure of my
professional interests is
on the author page.