Listen to this: Data and Analytics for Hiring and HR

From afar, I have to admit for some time I have been a fan of Evolv, an HR technology company that has for the last few years been been doing some really clever and interesting things in the assessment, screening and analytics space.

Essentially, Evolv helps big organizations, like Xerox for one, understand the characteristics, experiences, and skills that tend to make people successful (and not successful) in a given job, and then helps organizations test for and hopefully hire, the kinds of people that meet (or come close to) those characteristics and therefore are most likely to be successful if they are hired.

Evolv can then evaluate these data points, (and they have millions of them), compare the answers given by prospective candidates to the profile of characteristics of existing employees that indicate eventual performance success, and assess the 'match' of the candiate's answers to the characteristics of the people that time and history have proven to actually be successful on the job.

Recently the folks at NPR's Planet Money took a look at the online predictive assessment process that companies like Evolv are developing and their experience and observations make for a really interesting listen. Check out the NPR podcast here, or using the embedded player below, (email and RSS subscribers may need to click through)

Really fascinating discussion, and mostly because it gives a glimpse into what 'regular' folks, i.e. people not into HR tech or HR or tech, and rather the typical job seeker, think about this approach to employment assessment and HR technology.

The bottom line seems to be that while there might be some (initial) reservations about the relevance, accuracy, and applicability of these kinds of screening tools for employment, that the 'old' system of resumes, cover letters, interviewing polish, and 'Do you have a friend on the inside?' cronyism that are the hallmarks of the traditional job search are no better than casing your lot with an online assessment.

For what it is worth, I think the approach Evolv is taking represents the future of screening, assessment, and hiring. Wordsmithing resumes, spending hours and hours on cover letters than no one reads, and trying to decide if you should wait 24 or 48 hours to follow up with a recruiter after an interview seem incredibly nonsensical, add little value to the process, and ultimately have nothing at all inherent in them that can predict eventual success on the job.

So take a listen to the podcast (about 18 mnutes or so), and get an idea where the future of assessment and recruiting is heading. Let me know what you think.

Happy Geico Day!

REPRISE: Happiness and HR Data - Coming to a Delivery Truck Near You

Note: The blog is taking some well-deserved rest for the next two weeks (that is code for I am pretty much out of decent ideas, and I doubt most folks are spending their holidays reading blogs anyway), and will be re-running some of best, or at least most interesting posts from 2013. Maybe you missed these the first time around or maybe you didn't really miss them, but either way they are presented for your consideration. Thanks to everyone who stopped by in 2013!

If 2013 was the year of Robots and Automation, then the first runner up for topic of the year would probably have been Data and Analytics. The below post was my personal favorite example of the topic and what the future (the near future I bet) will hold for how data about people will be combined with data about machines and mashed up with process design in order to drive business outcomes. The piece originally ran in August 2013.

----------------------------------------------------------------------------------

Happiness and HR Data - Coming to a Delivery Truck Near You

Sometimes in all the conversation in the HR/talent space about the increased use of data, Big Data, and workforce analytics by HR leaders and organizations that practical, innovative (and possibly somewhat creepy), examples of how all this data coupled with better tools to understand it all are sometimes hard to find. Or hard to understand. Or not really specific enough that they resonate with many HR and Talent pros.

Lots of the articles and analysis about data and analytics for HR end up reading more like, 'This is going to be important', or 'This is going to be extremely important and you are not ready for it', or even 'This is going to be extremely important, you are not ready for it, but I (or my company) is ready to help you sort it out.'

Fortunately for you, this is not one of those kind of articles.

Over the weekend I read a long-ish piece called Unhappy Truckers and Other Algorithmic Problems on the Nautilus site, that provides one of the most interesting and practical examples of how a better understanding of HR data, (among other things), is helping transportation companies plan routes, assign work, and execute managerial interventions, often before they are even needed.

At the core of most transportation and delivery problems is essentially a logistics challenge as the 'Traveling Salesman' problem.  Given a fixed time period, say a day or an 8-Hour shift, and set number of destinations to visit to make sales calls, how then should the traveling salesman plan his route for the maximum efficiency. 

For a salesperson making four or five stops in a day the problem is usually not that hard to solve, but for say a UPS or FedEx delivery truck driver who may have as many as 150 stops in a day - well that problem of math and logistics gets much, much more complex.  And, as the piece from Nautilus describes, the Traveling Salesman problem is not only incredibly important for transportation companies to try and solve, it becomes even more complex when we factor in the the delivery drivers are actual human beings, and not just parts of an equation on a whiteboard.

Check out this excerpt from the piece to see how one (unnamed) delivery company is taking HR and workforce data, couples with the realization that indeed, people are a key element,  and baking it in to the classic math problem of the Traveling Salesman:

People are also emotional, and it turns out an unhappy truck driver can be trouble. Modern routing models incorporate whether a truck driver is happy or not—something he may not know about himself. For example, one major trucking company that declined to be named does “predictive analysis” on when drivers are at greater risk of being involved in a crash. Not only does the company have information on how the truck is being driven—speeding, hard-braking events, rapid lane changes—but on the life of the driver. “We actually have built into the model a number of indicators that could be surrogates for dissatisfaction,” said one employee familiar with the program.

This could be a change in a driver’s take-home pay, a life event like a death in the family or divorce, or something as subtle as a driver whose morning start time has been suddenly changed. The analysis takes into account everything the company’s engineers can think of, and then teases out which factors seem correlated to accident risk. Drivers who appear to be at highest risk are flagged. Then there are programs in place to ensure the driver’s manager will talk to a flagged driver.

In other words, the traveling salesman problem grows considerably more complex when you actually have to think about the happiness of the salesman. And, not only do you have to know when he’s unhappy, you have to know if your model might make him unhappy. Warren Powell, director of the Castle Laboratory at Princeton University’s Department of Operations Research and Financial Engineering, has optimized transportation companies from Netjets to Burlington Northern. He recalls how, at Yellow Freight company, “we were doing things with drivers—they said, you just can’t do that.” There were union rules, there was industry practice. Tractors can be stored anywhere, humans like to go home at night. “I said we’re going to need a file with 2,000 rules. Trucks are simple; drivers are complicated."

Did you catch all the HR/talent/workforce data baked into the model described above?

Payroll, time and attendance, life events that likely would show up in the benefits admin system, scheduling are all mentioned, and I bet digging deeper into the model we'd find even more 'talent' elements like supervisor or location changes, time since a driver's last compensation increase, and maybe even 'softer' items like participation in company events or number of unread emails in their inbox.

The specifics of what bits of talent data aere being incorporated into the process matter less than the fact that in the example the HR data is being mashed up so to speak with the 'hard' data from the truck itself (which is another interesting story as well), and analyzed against past driver experiences to alert managers as to when and where an accident is more likely to occur.

There is even more to the problem than the technical observations from the truck itself, and the alogorithms' assessment of the HR/Talent data - things like Union rules and contracts factor into the equation as well. 

But for me, this example of taking HR data and using it not just to try and 'predict' HR events like involuntary turnover or a better or worse performance review score, and apply it to real business outcomes, (the likelihood of accidents) represents a great example of where 'Big Data for HR' is heading.

I definitely recommend taking a few minutes this week to read the entire piece on the Nautilus site, and then think about some the next time the FedEx driver turns up with a package.

REPRISE: If Yahoo doesn't kill remote working, then Big Data will

Note: The blog is taking some well-deserved rest for the next two weeks (that is code for I am pretty much out of decent ideas, and I doubt most folks are spending their holidays reading blogs anyway), and will be re-running some of best, or at least most interesting posts from 2013. Maybe you missed these the first time around or maybe you didn't really miss them, but either way they are presented for your consideration. Thanks to everyone who stopped by in 2013!

The below post hits on a couple of topics that were beaten to death in the HR blogosphere in 2013 - the talent management decisions at Yahoo! and what technology and Big Data will mean for work and workplaces. The piece originally ran in March 2013.

----------------------------------------------------------------------------------

If Yahoo doesn't kill remote working, then Big Data will

A little bit lost in the continuing fallout from the decisions by Yahoo to end remote working arrangements for their staff, and Best Buy's move to end ROWE (Results Only Work Environment), at its corporate headquarters was this much more interesting, (and potentially more important), report in the Wall Street Journal, 'Tracking Sensors Invade the Workplace', that hints at a data-powered future workplace where 'being physically together' is not just mandated, but is tracked, recorded, and interpreted by algorithms and leveraged by management.

How exactly does Big Data, (which usually sounds kind of benign, or at least non-threatening), play a role in the future of telework?  Take a look at this excerpt from the WSJ piece:

As Big Data becomes a fixture of office life, companies are turning to tracking devices to gather real-time information on how teams of employees work and interact. Sensors, worn on lanyards or placed on office furniture, record how often staffers get up from their desks, consult other teams and hold meetings.

Businesses say the data offer otherwise hard-to-glean insights about how workers do their jobs, and are using the information to make changes large and small, ranging from the timing of coffee breaks to how work groups are composed, to spur collaboration and productivity.

"Surveys measure a point in time—what's happening right now with my emotions. [Sensors] measure actual behavior in an objective way,"

The next step in figuring out how people work, communicate, and interact in the workplace and with their colleagues involves wearing an always-on tracking device, (bathroom breaks optional), and harnessing all the data the device collects about who a worker talks to and for how long, how often they get up, when they hit the coffee room and vending machine, how long they stand waiting outside a conference room because the prior meeting ran long - all of this and more.  Mash up that 'experience' data with other electronic data trails (email, IM, internal collaboration tools, etc.), and boom - the data will be able to prescribe optimal amounts of employee interaction, recommend the timing and duration of breaks, send push notifications alerting you that the guy you need to connect with about the Penske account is two stalls away from you, and crucially - keep your managers informed about just what the heck you are up to all day.

But it seems really likely to me that if these workplace tracking sensors gain more well, traction, that organizations will quickly realize that the only way to really exploit them, and the data they collect to its fullest potential, will be in a traditional workplace environment - with all employees together in a physical location and 'on-duty' at the same time. Let's face it, for a remote worker wearing a tracking sensor probably won't produce much valuable data - unless its to try to 'prove' to a suspicious manager that a remote worker is slacking off.

The tracking sensors, if they catch on, will change the anti-telework argument from 'We need you to come in to the office so we can keep an eye on you' to 'We need you to come in to the office so we can track everything you do, say, touch, and feel all day.'

It's a brave new world out there my friends...

Uber, surge pricing, and data at work (at work)

The on-demand black car service Uber took quite a bit of flack over the weekend for implementing what is known as 'surge pricing' during a pretty nasty snowstorm in New York City. If you are not familiar with Uber, (and you should be because it really is an amazing service), the basics are pretty simple. Users use a smartphone app to summon a black car or equivalent that picks them up and then are taken to their desired destination. The entire payment transaction (including leaving the driver a star rating) is executed via the app, for prices (at least in my experience) ranging 15-20% more expensive than 'regular' taxi service.

But during times of extremely high demand for rides and low supply of on the road drivers (like on a Saturday night in a bad storm), Uber implements 'surge pricing', essentially increasing the cost of rides anywhere from 2 to even 6 or 7 times the normal fares in order to balance demand with supply. The ECON 101 logic is pretty simple - the increased prices (which users are warned about in advance of booking a ride) will serve to simultaneously reduce demand while increasing supply, as more drivers will be enticed to get out on the road in order to earn increased fees during the surge pricing period.

In addition to using basic pricing flexibility to manage and try and balance supply and demand, Uber also is attempting to mitigate the one really frustrating piece of the typical customer's experience, (I can attest to this one), which is the simple lack of availability of a car when you need/want one.

But the backlash from last weekend's surge pricing in NYC seemed pretty harsh as people took to Twitter to vent about their frustration with Uber for radically increasing their prices during a time of "crisis" in the city - it seems like there were scads of celebrities that were particularly peeved about having to pay what they felt like were exorbitant prices for transportation around town.

Putting aside the natural lack of sympathy I have for anyone complaining that their on-demand, door-to-door, black car service costs too much (on a Saturday night in the busiest city in America and during a snowstorm), I wanted to highlight this story as one of the very few that we see that showcases how data, technology, and the combination of the two are actually conspiring to benefit the front-line worker - in this case the Uber affiliated black car drivers.

Normal taxi drivers or even limo drivers might see a little extra in their pay rates for working a Saturday night, but certainly could not take advantage of the dramatic increase in demand for their services as the Uber drivers who braved the storm were able to realize.

Through a combination of new technology, absence of the pricing regulations imposed on traditional taxi services, more flexible labor rules, and most importantly, the presence of information of the increased demand, these Uber drivers were able to make better and hopefully, more informed, data-driven decisions about whether, where, and when to provide their services.

Most front-line workers never really get the exercise the kind of labor pricing power that we see in this example. Last Saturday night lots and lots of pretty well-off people wanted black car service on one of the worst weather nights of the year. The kind of night that most folks would rather stay home and stay warm, much less venture out into the cold and wet and storm to work for their normal pay.

Thanks to data and technology at least in this example, the Uber drivers who did venture out into the weather did a little better than most front-line workers.

It looks like they were paid what they deserved. Which is not always easy to say, both for black car drivers and for the celebrities they ferried up and down Manhattan last Saturday night.

PODCAST - #HRHappyHour 173 - How Data is Changing Recruiting

HR Happy Hour 173 - How Data is Changing Recruiting

Recorded Friday November 22, 2013

This week on the HR Happy Hour Show, Steve Boese and Trish McFarlane sat down with Eric Owski, VP of Product Strategy for Bright.com, for an interesting and informative conversation about how data, Big Data really, machine learning, and really sophisticated algorithms are helping organizations better understand the fit and potential for high performance of their candidates, and increasing the chances of making better hires while reducing the time and expense to make screening and hiring decisions.

It is still a challenge for many HR and Recruiting organizations simply to manage the volume of applicants that they are seeing for many positions, and to have the ability to spend the time and resources attempting to ensure they are engaging with (and hiring) the very best people that they can. The volume can often make expediency win out over making informed decisions, and in front-line, customer-facing roles this has the potential to cause pretty significant problems for the organization. 

Bright.com offers an approach and a solution that is based on millions of data points, informed by talent and recruiting professionals' input, and validated by the companies that have used the innovative 'Bright Score' to make more consistent and correct decisions about talent and potential.

You can listen to the show on the show page here, on iTunes, (just search in the podcasts section for 'HR HappyHour'), and using the widget player below, (email and RSS subscribers will need to click through). 

More Business Podcasts at Blog Talk Radio with Steve Boese and Trish McFarlane on BlogTalkRadio

It was a really informative and forward-looking kind of conversation that sheds some light on what new ideas and technologies promise to deliver to the talent management world now and more and more in the future.

Thanks to Eric and everyone at Bright for taking the time to share his insights!