Shared publicly  - 
The Hiring Algorithms are coming

On call center workers:
The software said that what does matter in a good call-center worker—one who won't quit before the company recoups its $5,000 investment in training—is personality. Data show that creative types tend to stick around for the necessary six months. Inquisitive people often don't.

I wrote on a LinkedIn post about this yesterday in response toa post by +Jeremy Haskell :

But while the final decision likely must come from an informed human brain, I expect that the potential pool of possibles to be curated by algorithms and based in many cases on new measures of skill, reputation, and influence.

I expect an economy that will run more and more on reputation metrics and skill matching algorithms.
Meg Tufano's profile photoJeff Jockisch's profile photoMichael R. Bernstein's profile photoRick Nimo's profile photo
So creative people aren't inquisitive? I think they are using 'creative' as a nice word for something else.
What is the average pay for a call center employee vs. other entry level opportunities?

What is the career path of this position?

Do the employers view this position as anything beyond a "churn" prone field?

Personally I see a problem with employers that approach such positions with a "herd of cattle" mentality and that such an approach tends to reinforce the problems that they are trying to solve.
Breaking News: Good leadership learns how to work with their employees on a personal level.
Some great points above, +Jesse Newhart +Andrew Coffman +David Green  

Metrics and thresholds can be dystopic and discriminatory and counter-productive.

But make sure you are also looking beyond the potential for misuse. Misuse of information happens every day.  In fact, chances are your hiring managers and HR departments are already considering the  name recognition value of your previous employer more than your work, your college degree more than your knowledge, your credit score more than your aspirations.

In a global market, in a big data world, in a universe where most hiring managers actually use many poor cues to pick employees, NOT using the data points you can gather about a potential employee would be stupid indeed.

The ugly truth is that we have to sort and cull data / opportunities / potential employees in SOME way.  I think the new ways will eventually be better than the old ways. Maybe they already are...
Resume algorithms are already a failure so not buying into this one quite yet.
I'll stick with the old school Neptotism style hiring, it has known advantages and disadvantages that can be easily gamed for the favor of both.
I think you are right on that one, +Bearman Cartoons.

Skill matching is a big problem.  I'm meeting this coming week with a acquaintance that works for a large job board to discuss skill taxonomies.  Interesting issues with natural language, synonyms, data collection, and the moving target of hot skills themselves.
That's the thing, +Andrew Coffman. Lots of bias in hiring.  
Ultimately it is easier to take the bias out of an algorithm.  

And if done right, it could mean vast efficiencies and a world that works much more like the meritocracy it should be, rather than the false one we live with today.
Just read an article that said that the issue with the resume process is all the work you need to do.  Some with all the parts and personality tests can take up to 30 minutes for one job.  The best candidates can be missed if they are just casually looking and don't want to put in the effort.  I know I have backed out of applying on several occasions because of personality tests and other hoops.
Intangibles drive the more technical skill sets that are the ubiquitous moving target, I'm fairly certain that for entry level positions and mundane career fields that hiring algorithms will be a boon.

For engineering, research and executive positions (among others) the algorithm may not produce the intended results.

For example, I am an AWS certified welding inspector with multiple ASNT NDE certifications and experience in multiple industrial and construction codes and applications. The biggest entry wedge into an interview process is my experience and certifications, from there it is mainly the human elements of the interview and the hard numbers of the job offer that determine whether I am offered a job and choose to accept it.
Holden: Describe in single words only the good things that come into your mind about... your mother.
Leon: My mother?
Holden: Yeah.
Leon: Let me tell you about my mother.
The article misuses the term "big data".
I wonder whether these algorithms will latch onto correlations such as race, ethnicity, gender, body modification, etc. in a similar way to humans, thus reinforcing biases.
Definitely a concern, +Michael Bernstein

But with the laser focus on the bottom line most corporations show, I think they are likely to value efficiency over bias.

Would be hard to design a system with more bias than single human decisionmakers working with limited data... 
Well, sure, but some of those biases are based on observable correlations, so the signal is there to be detected. Will these algorithms point the causation arrow in the right direction and disregard the signal, or go for the easy (and efficient) filter?
I kind of disagree, Jeff.  Corporations pay attention to other factors besides efficiency.  And there are other sources of bias besides individual decisionmakers.
Example from the end of the article: People who move more frequently "have a higher likelihood of leaving,"

Oh, look, people with less financial stability in their lives will now have a harder time finding a job! Yay!
It's almost a tautology that people who move more frequently have a greater probability of leaving.  And it's true, many employers seek a stable workforce.  But even now, frequent job-hoppers are often screened out.
I'd agree that job history is fair game, but moving your residence shouldn't be a black mark. There is a high risk of a self-fulfilling prophecy here.
If I'm not mistaken, frequent moving also damages your credit rating, making it more expensive to borrow money.
Loan officers also make personality assessments before making loans, at least at some banks.
I would be scared of humans applying observable 'correlations' +Michael Bernstein that sounds like how biased decisions get made....

The financial stability issue you and +Rick Nimo mention is a good one. The building and applying of algorithms is indeed tricky.

Still, overall, I think new reputation metrics are likely to reward performance in ways our current system does not. In that way, it will empower job seekers with new ways to show worth, rather than relying on current reputation metrics that are undoubtedly biased and inefficient. 
Humans applying observable correlations are indeed how biased decisions get made, mostly because humans are very bad at choosing which correlations to pay attention to, and at weighting them appropriately.

That doesn't mean that just because algorithms based on big data will tend to apply those correlations correctly, that the result will be desireable.

The reason you can't ask people directly about their age, whether they have kids, are married, etc. has nothing to do with whether the correlation between those factors and undesireable-to-employers traits (example: single mothers typically needing more schedule flexibility to take care of kids and using sick days as a result) is statistically valid, and everything to do with the fact that filtering people out from employment opportunities based on those factors is just wrong.

Using Big Data powered algorithms to find better and narrower proxies for those traits (so that you're now discriminating against 'people who are likely to use sick-days when they aren't actually themselves sick' regardless of whether they happen to be single mothers), though legally defensible, doesn't seem to be a huge improvement to me.

Just because we can figure out more efficient ways to discriminate than this: doesn't mean we should.
I agree that such algorithms might provide a mechanism to justify discrimination, +Michael Bernstein

My contention is that this discrimination is already happening. And more data and new metrics are much more likely to help than to hurt.
What would happen if everyone in an organization had to work the call center an hour a day.  Just wondering if the problem is that there is not a new way to set up the job so that you have zero problems in keeping things running.  For example, I don't know anything about the question the client is asking about (her computer doesn't work (for a stupid example)); but I know what I've done in the past and I suggest it to her.  It happens to work.  Problem solved.  Or it doesn't work; and I know someone who can call her back when it's HIS hour who is an expert that can fix it in no time!  (Also called a network of informed people.)  Now, if you have a stovepiped company, that's not going to work.  But the point of "social" is not only about social media, it's about getting to know the people in your group WITHIN your company.

You tell me, any job you've ever had, wasn't it the people who made it great or horrible?  

"Entry-level" jobs are called that because they're often jobs that are boring, require all your attention, and usually require you to be over-the-top nice to people who are over-the-top not.  (Thinking "receptionist" (which doesn't much exist anymore), but also "help desks," "assistant," etc.)

I cannot see any reason why everyone cannot do some of these jobs some part of the day and to hell with requiring a person to spend their entire day doing something boring AND thoughtful (weird combination, but just try being a secretary (which I don't think they have anymore either)).  A really good secretary (now called an Executive Assistant) is one of the most creative, demanding jobs on earth.  You have to write what you think your boss wants to say to someone else.  You have to understand all the interactions going on in a big meeting so that you can write a summary of it in such a way that everyone signs off on it as accurate.  You have to get the dates right (back in my day, you had to actually write the airplane tickets), get the words right, the people right, the paper right and on and on and on.  (What they call "detail oriented.")  

Just saying, maybe our problem is not that we don't have ways of finding people who will "stick it out."  We have a problem in thinking back-in-the-day management is going to work for people who are now-in-the-day highly educated and desire varied occupations.  

Why does this problem persist?  Ask a CEO if he'd spend an hour a day answering calls.  Trust me, he or she would learn more in those five hours a week than in a Webinair (Man, I hate those things), but it's not sexy!  And it's just not "done."  

This is not actually a Big Data issue; I think it's a management ego issue.  (Same with college administrators; they should be required to teach one class a year.)
You are pointing toward a different problem, +Meg Tufano though perhaps one closer to the root.

I think that your suggestion, making business social, could solve a host of problems and improve efficiency.

And I also think that data and algorithms could help solve them. We can use them just as well after we hire employees... 
Well. As always, the real question is who has the data and the algorithms, and thus gets access to justifications for their choices.

Keep in mind that these tools could have been used instead to help people discriminate among companies and jobs to find ones they would do well in, but that isn't what we're seeing, since the readily accessible concentration of money and power lies elsewhere.

My point isn't that these tools will be abused to justify human prejudices, but that giving corporations and other large organizations more power to be efficiently discriminating will have the inevitable side effect of denying economic opportunity to those who most need it, end up reinforcing class (and race) divides, and reduce social mobility.

And it isn't just employment. We've already seen what insurance companies can do to reduce their liabilities and deny coverage to their most expensive customers, for example.
You make a great point about the insurance companies using data against us, +Michael Bernstein. I expect that lenders can and will do the same. Risk assessment industries might be problematic within my overall supposition that reputation metrics enable social mobility.

And the ownership of the data is indeed key, and a big problem.  I'm with you on that point and am an advocate of personal data stores and rights to privacy.

But I think the battle for privacy is lost already.  The best we can hope for is extreme transparency; requiring the uses of data to be exposed.

On the abuse of data to maintain and reinforce prejudice and class structure, I still disagree in general.  There will be cases of abuse, maybe even whole industries, but ultimately I think data will expose stupidity as being inefficient. 
This problem won't be confined to particular (risk-assessment) industries. If you re-read the article you shared, you can see that these companies are treating hiring as a risk assessment problem.

This also isn't a question of privacy. I can accept that the data is already out there, and will continue to pile up. The question is what are organizations, with their greater resources than most individuals, allowed to do with that information. And transparency isn't much of a help, if all it does is show that employers have perfectly valid and justifiable reasons for being discriminating.

As for prejudice being stupid and therefore inefficient, I am apparently not getting across my point that I am not particularly worried about deliberate use of these tools for prejudicial reasons, but that reinforcing race, and especially class, divides will happen as a side effect, in the name of efficiency.

Sure, we'll eventually also see data that organizational cultures that are too homogenous, personality wise, are brittle and less able to adapt over the long-term, but that won't stop the drive for short-term efficiency, just as it hasn't stopped agricultural monocultures, and meanwhile the cost in stifled human potential will be immense.
It's important to hire people who can get along with the people who are already working at a company.   If the factors that make people compatible are unmeasurable (or excluded from the hiring algorithm), then you're going to have a problem because you will be changing the hiring criteria, and the new hires will be different.  Although some people may think it is wrong,  companies do need to maintain consideration of short-term efficiency or they may not be able to survive until the long term.  And as Michael Bernstein points out,  the employment relationship is bilateral:  new hires also have preferences about where and with whom they wish to work.  Those preferences affect employment data, too.
Add a comment...