Learn a new word: Positive Predictive Value

Predictive analytics and the application of algorithms to help make 'people' decisions in organizations has been a subject of development and discussion for several years now. The most common applications of predictive tech in HR have been to assess and rank candidates for a given job opening, to estimate an individual employee's flight risk, and to attempt to identify those employees with high potential or are likely to become high performers.

These kinds of use cases and others and the technologies that enable them present HR and business leaders with new and really powerful tools and capabilities that can, if applied intelligently, provide a competitive edge realized from the better matching, hiring, and deploying of talent.

But you probably know all this, if you are reading an HR Tech blog anyway, and perhaps you are already applying predictive HR tech in your organization today. But there is another side or aspect of prediction algorithms that perhaps you have not considered, and I admit I have not really either - namely how often these predictive tools are wrong, and somewhat related, how we want to guide these tools to better understand how they can be wrong.

All that takes us to today's Learn a new word - 'Positive Predictive Value (PPV)'

From our pals at Wikipedie:

The positive and negative predictive values (PPV and NPV respectively) are the proportions of positive and negative results in statistics and diagnostic tests that are true positive and true negative results, respectively. The PPV and NPV describe the performance of a diagnostic test or other statistical measure. A high result can be interpreted as indicating the accuracy of such a statistic.

A good way to think about PPV and NPV is using the example of an algorithm called COMPAS which attempts to predict the likelihood that a convicted criminal is likely to become a repeat offender, and has been used in some instances by sentencing judges when considering how harshly or leniently to sentence a given criminal.

The strength of a tool like COMPAS is that when accurate, it can indicate to the judge to give a longer sentence to a convict that is highly likely to be a repeat offender, and perhaps be more lenient on an offender that the algorithm assesses to be less likely to repeat their crimes once released.

But the opposite, of course is also true. If COMPAS 'misses', and it sometimes does, then it can lead judges to give longer sentences to the wrong offenders, and shorter sentences to offenders who end up repeating their bad behaviors. And here is where PPV really comes into play.

Because algorithms like the ones used to create COMPAS, and perhaps the ones that your new HR technology uses to 'predict' the best candidates for a job opening, tend to be more or less wrong, (when they are wrong), in one direction. Either they generate too many 'matches', i.e., recommend too many candidates as likely 'good hires' for a role, including some who really are not good matches at all. Or they produce too many false negatives, i.e. they screen out too many candidates, including some that would indeed be good fits for the role and good hires.

Back to our Learn a new word - Positive Predictive Value. A high PPV result for the candidate matching algorithm indicates that a high number of the positives, or matches, are indeed matches. In other words, there are not many 'bad matches' and you can in theory trust the algorithm to help guide your hiring decisions. But, and this can be a big but, a high PPV can often produce a high negative predictive value, or NPV.

The logic is fairly straightforward. If the algorithm is tuned to ensure that any positives are highly likely to truly be positives, then fewer positives will be generated, and more of the negatives, (the candidates you never call or interview), may have indeed been actual positives, or good candidates after all.

Whether it is a predictive tool that the judge may use when sentencing, or one your hiring managers may use when deciding who to interview, it is important to keep this balance between false positives and incorrect negatives in mind.

Think of it this way - would you rather have a few more candidates than you may need get screened 'in' to the process, or have a few that should be 'in' get screened 'out', because you want the PPV to be as high as possible?

There are good arguments for both sides I think, but the more important point is that we think about the problem in the first place. And that we push back on any provider of predictive HR technology to talk about their approach to PPV and NPV when they design their systems.

PODCAST - #HRHappyHour 270: There are no new HR predictions

HR Happy Hour 270- There Are No New HR Predictions

Hosts: Steve Boese, Trish McFarlane

Listen to the show HERE

This week on the show, Trish and Steve discuss why there are no new HR predictions.  Every year by January, there are a slew of articles, blog posts and social media discussions around HR predictions.  As usual, we're finding that none of the predictions are new.  They tend to be restated from previous years. This episode talks about that and what business leaders should be focused on instead.

We also spent some time talking about what is coming up in 2017.

You can listen to the show on the show page HERE, or by using the widget player below (email and RSS subscribers need to click through)

This was a fun and informative show, we hope you like it! Thanks to show sponsor Virgin Pulse -learn more about them at www.virginpulse.com.

Remember: Subscribe to the HR Happy Hour on iTunes, Stitcher Radio, or your favorite podcast app - just search for 'HR Happy Hour' to find and subscribe to the show.

Thanks to everyone who listened in 2016 and Happy HR New Year!

Onboarding the algorithms

During the Ideas and Innovators session at the HR Technology Conference last month, my pal Michael Krupa gave an outstanding talk about automation and advanced 'learning' kinds of technology, and some of the implications for HR and organizational leaders who are choosing to incorporate these technologies into their people and talent management programs.

In the talk, (and once I get the video of this, I will be sure to update the post with the link), Michael used a great expression to illustrate the deliberate and measured approach HR should take to adoption of these 'smart' tools. He called it 'Onboarding the algorithms'; a way of comparing the introduction and deployment of these technologies to the structured and immersive process that most organizations follow when onboarding new employees. The larger point - HR and business leaders need to carefully evaluate, understand, assess, and introduce algorithms and other advanced, intelligent, (and often predictive), tools carefully, and insert them into the HR and talent processes intelligently and intentionally - just like we do when hiring and welcoming new employees.

I think Michael's analogy was a great one as it serves as a kind a warning to HR and business leaders eager to adopt these kinds of advanced and predictive tools for functions like evaluating a slate of job candidates, making decisions about which employees should be considered 'high potential', and thus granted more development and growth opportunities, scheduling the 'optimal' mix of employees for a given day or shift, or to provide intelligence and decision support to managers making decisions about the allocation of salary and bonus pools.

I have no doubt that organizations and HR leaders will seek to adopt these kinds of tools more and more in 2017, but at the same time I think it also we be important, to use Michael's phrase, to 'onboard' these tools and algorithms effectively, in order to ensure we not only utilize the tools to their potential, but we also understand how these tools are actually designed and how they are performing. Kind of like how we know the background of every new employee that comes onboard the organization and how we like to keep track of their assimilation and performance - usually in a pretty structured 30-60-90-180 day kind of manner.

This is a pretty important and complex idea for sure. One that can't be completely explained in one blog post. But I think one good place for HR and business leaders to start is to have really open and honest conversations with their HR technology providers of these smart tools and algorithms to gain a better understanding of how they are designed, how they work, (or are meant to work), and how transparent are the machine's thought processes they end up in a decision, (or at least a recommendation, i.e. 'Hire this person, and not that person.').

A great starting point for HR leaders who need to know what questions to ask of your HR tech providers is the Principles for Accountable Algorithms statement from the  Fairness, Accountability, and Transparency in Machine Learning organization. They break down the five key guiding principles for algorithm design, (responsibility, explainability, accuracy, audibility, and fairness), that HR leaders can look for, (and seek answers about), from their technology providers. Of course there is plenty more for HR leaders to consider when evaluating and deploying these tools, but the FAT/ML document provides a good starting point. You can learn more about that organization and what they do here. 

Just like every good HR pro knows that we just can't toss new employees blindly into the fire and expect that they will produce desired outcomes, (and be happy), we can't expect that we can simply insert new technologies, no matter how 'intelligent', and think we will get optimal outcomes. If these tools are meant to become a fundamental element of you and your leader's talent and people management playbook, then you need to understand them as well as you need to understand all the members of your team.

And you need to be able to tell when the algorithm is wrong too.

Have a great week!

The former next big thing

I was doing some cleaning/organizing/trying to find something (there's almost no difference), over the weekend when in a box of old papers and books I came across the CD that you see pictured on the right (click on the pic for a much larger version).Click for an enormous version of this

Since it still might be a little tough to catch all the details from the pic, let me explain what you're looking at.

This CD dates from 1997 and contains the Functional Overviews, (essentially user manuals), for Oracle Applications products for Finance, Supply Chain, Human Resources, and more. But more notable than the fact that I still have an almost 20-year old CD set of Oracle Apps manuals, is the particular Applications release or version number of the products - something called 10SC.

For those reading this post not familiar with Oracle Apps Release history the 'SC' in 10SC stood for 'Smart Client' - in Oracle-speak, 10SC was the initial version of the applications that were deployed using graphical forms instead of character-based input screens, and in true client-server mode.

After typing that I just realized there are probably lots of readers that in addition to not being familiar with old Oracle Apps version numbers, probably are not even that familiar with the notion of 'client-server' as well. Put really simply, client-server was/is a method of deploying applications installed on a central computer (the server) to the end users (the clients) over a local area network, a wide area network, or, later, the internet.

Prior to the client-server revolution, business applications were primarily installed on mainframes and minicomputers that possessed input/output terminals, (I know, this post has officially become incredibly boring, I promise I am getting to the point soon), so the migration to more advanced client-server application architectures was a VERY BIG DEAL.

I mean a big, huge, massive deal. If we had Twitter or LinkedIn back then you would have been inundated with pieces like '10 Ways Client-Server is Going to Change How Work Gets Done' and 'Client-Server offers HR Leaders the Opportunity to Re-shape HR service delivery'. And if you were not on board with Client-server for some reason, you'd have been labeled some kind of technology Luddite and a 'typical HR person' getting in the way of progress because you didn't stay on top of technology trends.

Kind of the same way that the 'Cloud' has been referred to in about 14,980 articles written this year.

So here is the point. It is not that the Cloud and how HR technology has changed dramatically in the last few years isn't important. It is. It does matter and I firmly believe that owning the HR and workplace technology agenda is the true secret to developing lasting influence and power in organizations.

It is just that this has pretty much ALWAYS been the case and the Cloud or smartphones or wearable devices are just the latest manifestations or versions of Client-server, or Email or Fax machines.

Sure you might be tempted to say 'You are wrong, things are moving faster. It is different this time.'

You are right. It is different this time. It's different EVERY time.

Keep this in mind as you wind down 2014 and read any 'Top HR and HR Tech predictions for 2015' pieces. Even ones where your humble correspondent might be quoted.

Happy Thursday.

Note: If anyone from Oracle is reading this and wants the 10SC CD, let me know I will send!

Ten things to watch in 2014

At the start of the New Year the marketing/branding/digital design firm JWT releases a really cool and interesting collection of '100 Things to Watch' for the upcoming year - a collection of new ideas, trends, technologies, societal shifts, etc. that are meant to stimulate thinking and generate discussion. Many of the 'things to watch' are kind of uber-trendy and destined to be largely unimportant and fleeting, (fast food tofu, Oculus Rift, and Chinese Wine), but others, particularly the tech trends that JWT identifies have the potential to be more significant, enduring, and even influential in the design of workplaces and the nature and manner in which work is performed. And last year we even had some fun talking about some of these items, 'Adult Playgrounds' in particular, on the HR Happy Hour Show in the past.

The entire JWT slide presentation is embedded below, (email and RSS subscribers may need to click through), and after the deck I'll pull a few of the 'things to watch' that are likely to be more relevant and meaningful to work and workplaces.

JWT: 100 Things to Watch in 2014 from JWTIntelligence

So which of the 100 things should you as an HR and Talent pro be on the watch for?

Here are just a few from the JWT deck that I think you should keep an eye on.

6. Ambient Commerce - As a means to improve customer service, 'ambient' systems will use data and signals to track consumer needs, anticipate them in advance, and provide products and services before they are even requested. This is the 'smart refrigerator' that orders more milk for you one day before it thinks you are going to run out (or it spoils). For our workplaces, meeting this expectation of our customers is going to require us to spend more time and resources on new systems and methods to better understand what our customers are likely to want to do next based on what they have done in the past.

8. Arrested IRL Development - Growing up online and attached to your smartphone could mean an entire generation of new workers with less or even limited understanding of how to work, communicate, and socialize in real life (IRL). For HR and the workplace this could mean more time and attention spent in new hire onboarding programs on things like running and participating in meetings, writing and communicating more formally, and even how to handle yourself at a business meal or corporate cocktail hour.

11. Beacons - These are the signals that network-connected smartphones send and that are going to be increasingly used to identify, track, and customize the on-site shopping experience for customers at all kinds of venues. Retail store, museums, and even sporting events are just a few of the potential applications of this technology. For HR and work the implications are vast - but the first and most likely could be in retail or other high customer touchpoint businesses. Workers can be given much more specific, targeted information about customers and prospects prior to making initial contact, and will be required to process and evaluate much more information prior to engaging the customer as well.

25. Contemplative Computing - This is a big, catch-many things kind of idea, but the important element is meant to answer the question of 'How can our computing/devices/social networks become less of a distraction and burden and more of a helper, guide, and trusted advisor?' For work and workplaces, it is almost certain that many of our employees are already feeling significant information overload, (that is getting worse by the day), so smart organizations will take steps to ensure the systems and devices that employees use are not placing even more stress on them, and rather, are actually helping them get their jobs done, (which will make them less stressed out).

35. E-cigarette Regulation - This could be a big 'nothing to see here' kind of trend, as it can be really simple, especially for workplaces, to interpret and treat E-cigarettes the same way most of us treat tobacco products already (banned at work, we won't hire you if you use them, if you do use them be prepared to pay 476% more for your company-sponsored medical insurance, etc.). But if the E-cigarette market continues to grow, and employees and customers push back on rules and regulations that restrict their usage, then HR will need to sort out the best way forward. Or you could just pro-actively ban them tomorrow and not worry about this one.

37. Equal rights for men - This one might provoke some calls of BS from readers, as the workplace has pretty much always been a man's world, so rather than try and convince you on this, I will just pull the quote from the JWT slide and let you make your own determination if or when this trend might impact your organization. 

According to our research, men feel it has become harder to be a man today, and harder to succeed in the working world as a father and husband. Watch for a rise in male-focused support systems and advocacy groups as society comes to understand that many me would be well-served by some of the mechanisms in place to boost women.

41. Glanceable User Interfaces - Think about the best, most helpful, and easiest to use apps you have on your phone right now. Chances are they provide just the right amount of information to make a decision or get a needed update in just a glance or a swipe. So much of how we process, take in, and interpret signals and information is happening on tiny screens that we operate with one thumb that designers and developers are challenged to make what is presented on the first glance be meaningful and relevant. For you, it means one thing primarily - distill your delivery of information as far down as possible, and make it work on a 4-inch screen. Unlike this post, which is going to clock in at over 1500 words I bet.

47. Hashtag fatigue - What once was a clever way to help organize and make findable related information on social networks has morphed in many instances into an incredibly annoying game of 'Look at me and look at my stuff'. If you are a heavy Twitter user you know what I am talking about - people and brands 'hashtagging' unrelated (or even every single update) content with a popular or trending hashtag in a desperate and pathetic attempt to be noticed. If you do this personally or as a part of your HR/Recruiting game, you should stop. #Imeanit.

50. Human Touches - This is the natural backlash to an increasingly technology augmented and disconnected world. JWT sees opportunity for organizations that can figure out ways to supplement their technological interactions with customers (or for HR, employees or candidates), with personal and more human touches. This is not going to be easy for most, as the same technologies that enable processes to succeed at scale simultaneously work against the ability or capacity to allow more human contact. If you can figure our the best way to balance these conflicts however, you will likely have an edge in the markets in which you compete.

63. Minute to read it - In our increasingly busy, over-stimulated, and time-crunched world accurate estimates for the length of time it will take to read a report, process a transaction, or even sit in a meeting or phone call are becoming more and more important. How much more could you get done if you could only be more precise about how long it will take to actually get something done? For HR and Talent - maybe pushing back on your HR Tech providers to build this kind of predictive estimation into their tools would be a start. Think of annual benefits open enrollment or processing a self-appraisal. Giving employees and up-front (and smart, as it should be a learning algorithm) estimated length of time to complete would be a real win in most organizations.

That was 10 things that I thought were interesting and/or had some potentially relevant implications for HR and Talent pros, but looking through the entire list I could not stop at 10, so here are two more bonus trends to consider in 2014:

72. Robotic Security Guards - You know I could not resist shoe horning some Robot content into a piece like this, right? To me, this one is kind of a no-brainer. The technology already exists, the demand for this kind of solution seems likely to be significant, and it could even be a way to slowly introduce 'companion worker' type robots into more settings. In 2014 you might just have your first interaction with one of these.

97. Verified Reviewers - Ever wonder if those Yelp or TripAdvisor reviews that you rely on to make restaurant or travel decisions are actually real? Or if they are being left my people with similar tastes and sensibilities as yours? Well this trend, similar to the traditional old school newspaper movie or food critics, points to a future where the usually anonymous online reviewer ascends to a more verifiable (and possibly influential) position. What might that mean for your HR or talent role? Probably something that points to the importance of strong and trustworthy brand advocates, be they external to the organization, (thinking former employees most likely), or internal ones (that have developed a reputation for integrity that moves beyond company or brand shills).

So that's it - another epic take on the upcoming trends for next year, (I am sure you are sick of them).

What do you think? On the mark? Crazy?

Time will tell.

If you take a few minutes to have a look through the entire list, let me know what other 'Things to Watch in 2014' you think will impact the world of work.