Receive daily articles & headlines each day in your inbox with your free ERE Daily Subscription.

Not logged in. [log in or register]

The Big Data HR Fad

by Jun 19, 2013, 6:45 am ET

Screen Shot 2013-06-12 at 9.08.04 AMNothing excites organizations like another fad. The latest one happens to be a thing called “Big Data.” Big Data refers to collecting so many performance numbers that understanding them becomes difficult. Some people suggest Big Data be applied to HR, which brings me to my point. While Big Data might work for managing things and numbers, how can it apply to something few understand, let alone manage and measure … like human performance?

Human performance is A + B = C … that is, something stimulates the employee/manager (A), he/she does X or says Y (B), and the result is either good or bad (C). For example, a manager might have two problem employees (A), he/she talks to them (B), and later, everything is all better (C).

Sound simple? Sure, we can often record results (C), and sometimes we can even record the problem (A), but what the heck happened in the middle? Shouting? Warning? Exploring differences? Coffee chats? Bribery? Threats? Blackmail? Extortion? Something else?

Human performance is all about B … not A or C. That is, we need to know the specific employee skills used. Pile all the before or after-the-fact performance data you can collect into one big database and it still won’t be actionable until you include links to employee skills.

Skills are the things people bring to work. It’s the thing(s) they use to get the job done. It might be the ability to learn new skills, acquire specific technical knowledge, analyze data, make good decisions, be organized and able to plan, be motivated to act in a specific way, be skilled communicating with people, or any one of dozens of other job-related KSAs. If you promote an individual contributor to manager and the person fails, it’s probably because the “B’s” for the old job did not match the “B’s” for the new one.

Organizations are great soothsayers. They think reading the tea leaves of results leads directly to employee skills. I’ve known salesmen, managers, and business owners who were rewarded for performance, but the unethical practices used to get them there almost destroyed the company and workforce. You see, if you only have performance data, you never know the full story: did the employee lie, cheat, and steal; be at the right place at the right time; take credit for someone else’s work; or, were the results influenced by something else?

Big Data analysis might give you a fuzzy sense of confidence, but unless you understand your ABCs and include them in your data files, Big Data will be just another short-term HR fad.

This article is provided for informational purposes only and is not intended to offer specific legal advice. You should consult your legal counsel regarding any threatened or pending litigation.

  • Pingback: | Are You Losing Hires Because of Fat Bias? + MORE | Get Hiring

  • Richard Melrose

    Wendell,

    Thank you for this. Regardless of the field of business practice, “Big Data” has become the new “Blah, blah, blah.”

    People with little knowledge of statistics, and even less knowledge of human behavior tout Big Data, as a panacea. It seems to come up at every turn, these days.

    Yet, whether employee performance, customer utility or buyer behavior, one must acknowledge a complex fabric of “little reactions” – i.e. distinctly individual choices made in response to individual circumstances, desires and perceived alternatives. Moreover, people do not use deterministic, rational, time-stationary, commonly held,[computer-like] decision-making mechanisms … they choose for themselves, at instances, without necessarily revealing any of the perceptions, motivations and/or feelings behind their choices.

    To aggregate such “little reactions” and algorithmically crunch the resulting “Big Data” constitutes folly, both statistically and behaviorally. The father of the computer, Charles Babbage, understood this writing (in 1864): “On two occasions I have been asked, ‘Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?’ … I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.”

    I don’t get it either, Chuck. What confusion of ideas generates such [superficial, mindless] excitement about Big Data?

  • Richard Araujo

    “While Big Data might work for managing things and numbers, how can it apply to something few understand, let alone manage and measure … like human performance?”

    That’s an excellent question, and I think another problem that may over ride not knowing the “B” is that many people either can’t or won’t define performance. They’d rather make decisions from their gut based on whether or not they like this person or that, or if they seem like a “go getter!,” or any number of a billion other buzz words or catch phrases designed to communicate desirability, or the lack thereof, while being completely vague and imprecise as to the underlying reasons.

    As far as I’m concerned every job breaks down to a deliverable with a time frame and a quality standard. There may be subjective elements but in the end you still have to produce something at a certain time and to a certain standard. The skills are what’s needed to achieve that deliverable, and are in my opinion often way less specific than most people insist. They usually boil down to smarts and being able to get along with the people you work with, and wanting to be there, doing the job. Of course the more specialized you get in terms of jobs, the more specific skill sets and education matter.

    I couldn’t agree more about the potential for big data being a fad, though I hope it goes in the opposite direction. I think it was Dr. Sullivan who wrote another article here on data/metrics breaking it down into historical, real-time, and predictive. Unfortunately many people tend to think the historic and real-time are, or preempt, the predictive. To get the predictive data you need to have a theoretical framework, at least a guess as to what “B” is, in which to analyze the historical and real-time data so you can see if your theory measures up. And you have to be willing to make decisions based on that theory and eat the cost. That means risk, which people are averse to, which is why falling back on the excuse of, “this has always worked in the past,” is what happens so much of the time.

  • Richard Araujo

    “I don’t get it either, Chuck. What confusion of ideas generates such [superficial, mindless] excitement about Big Data?”

    Knowledge is power, and data is the illusion of knowledge for many people who don’t realize that having more numbers to crunch doesn’t let you predict the future. As such, I will continue to rely on astrology for now. Most everyone is aware at this point that it’s BS, so I can at least have fun with it.

  • Keith Halperin

    @ Everybody: The point of most such recruiting fads is (as I often mention) for slick hucksters with high-level connections ready to sell “recruiting snake oil” aka, “the latest recruiting fad” to desperate and not-yet insolvent recruiters and their superiors who fail to recognize that in most cases they are futilely “rearranging the deckchairs on the Titanic” of their companies’ ill-conceived, over-blown, grossly-dysfunctional hiring practices.

    The hucksters do this by appealing to two very strong human desires:
    1) The desire for control over the uncontrollable.
    2) The desire for simple answers to complicated and complex problems.

    IMHO, almost every recruiting fad I’m familiar with has/had some useful and valuable potential, which has usually been generalized, diluted, and/or over commercialized into frequent near-uselessness.

    Cheers,

    Keith

  • http://www.ets.org Michael Krause

    I’ve seen quite a few articles about metrics, data, productivity measurement, etc. and I think the last sentence of this article speaks volumes. Anything worth doing is worth measuring. How do you know you’re doing a good job if you don’t measure? How do you know how to improve your process if you don’t measure and gather data? Data is important but as the last sentence states, if you don’t understand what the data means, where it comes from, how to interpret it or even how to utilize it to make improvements, it is worthless.

  • Keith Halperin

    @ Michael: “Anything worth doing is worth measuring.”
    There are a lot of things which are worthwhile but hard/impossible to measure, but many of them are irrelevant to business. It’s important to realize these things:

    1) The data is the means to the end, not the end itself.
    2) Just because you can collect and analyze (C & A) data, doesn’t mean you should- if the (C & A) costs exceed the benefits in doing so, then you probably shouldn’t (C & A) the data.
    3) You shouldn’t have recruiters spend more than about 5% of our time (C & A) data- it’s not worth $50+/hr to do that.

    Cheers,

    Keith “Dammit Jim, I’m a Recruiter, not a Bit Diddler” Halperin

  • http://www.linkedin.com/in/shanil/ Shanil Kaderali

    Dr Williams. I enjoyed your article. Agree on some points but differ on others.

    I’m not ready to dismiss big data for HR as a fad yet.

    “While Big Data might work for managing things and numbers, how can it apply to something few understand, let alone manage and measure … like human performance”

    There is truth in your comment and I think Richard is right too. Reality is complex, uncontrollable and this may very well be that attempt to control the uncontrollable which is part of human history.

    However, also much the way that Wall Street quants (from places like MIT, Chicago and Berkeley) transformed high finance, the HR big data geeks have just begun and I don’t know if they’ll be transforming human performance analytics meaningfully, but they’ll try and right now, big venture capital/private equity is putting money into it. There’s never been this much data to even come near to the level of predictive analytics seen prior.

    For front-end process of recruitment, it will mean less recruiters in the sourcing and pre-qualification stages.

    Will big data lead to moneyball (the book) type of human performance analytics for companies in all types of roles? maybe

    In the employee review stages, what it does remains to be seen.

  • Srikant Raghavachari

    Williams, big data is not about the accumulation of huge performance numbers that is being captured that confuses people as to what they mean. It is about interpreting those data and analyzing them to make meaningful insights to it. Big data as you expressed was always available in different forms and structures stored in different media, but it was difficult to process these in a timely manner. The current hype is about the availability of technology that could make providing these insights in a timely manner for decision making.

    As an example

    Traditional screening of resumes involved looking for some keywords based on what the candidate has entered in his resume. Now the linkedin’s model of having your peers and managers endorsing your skills and providing recommendation weighs more on the skills you really possess. Is that 100% accurate ? no but it is an evolution that is now made possible because of BIG DATA and the results are going to get only better.

    Big data could facilitate drawing the insights about a person based on his email communications with his team. That could identify the skill he used to manage a specific situation.

    Whether it is being utilized properly in a specific industry is a question that only time will answer but there is a huge potential if HR focuses on identifying the key metrics and facilitates collection of those data points.

  • Richard Araujo

    @ Srikant,

    I don’t think anyone objects to using the data, in fact I think most people want what you describe. It’s just an unrealistic expectation because people are people, and being people, they screw up a lot. There will likely be some good things to come out of big data as you describe it, but it will also be used as a crutch and an excuse by those who want crutches and excuses. “Typically people with X background don’t work well in this company, so we shouldn’t hire this candidate because of X in his history…”

    The quants brought up by Shanil are a perfect example, because it’s largely these dipsticks and the people they enabled that created the environment, and the data, which convinced a lot of people that lending boatloads of money to people with bad credit and no hope of paying it back was a good idea, an economic stimulus if you will. Now this isn’t the place for an in depth discussion of what lead to the financial crisis, but the quants and their precious models did play a massive role. The problem with them specifically is they started believing in the data as if it were the theory. In complex systems where few if any controls are possible it’s impossible to truly conduct experiments in a scientific way.

    One of the ways the traditionally soft sciences like economics and sociology deal with this is with mathematical/statistical models. The problem is that even when these models do achieve a fair level of accuracy in their predictions, because there has been no truly controlled study involved, what they’re really doing is playing the odds based on historical data and trends, and correlative relationships which over time are assumed to be causative, until the model breaks down, and then it gets tweaked until it more or less agrees with historical data and the process starts over again.

    I don’t know that big data isn’t moving in this general direction. I hope it doesn’t, but my gut tells me it likely will.

  • http://www.ScientificSelection.com Dr. Wendell Williams, MBA, Ph.D

    @Srikant…your example illustrates my point…a manager’s email content is a clear example of communication style. Assuming you could accurately evaluate it, it’s the “B” in the ABC equation. But your result (C) is unclear (i.e., what does “manage a specific situation” mean?

    Now let’s take it a step further, how will BD explain which specific employee skills lead to balancing a budget? Or implementing an employee intervention program? Or reducing turnover? Working only with results (or what BD usually collects as data) is only partial information.

  • Keith Halperin

    @ Everybody: There’s a great deal of discussion of “Big Data” in general, but very little about what specific things it would be used to show. What would YOU like to use it for” (I don’t mean things like “being able to source or hire better candidates”- please be more granular.)

    Cheers,

    Keith

  • Richard Araujo

    @ Keith,

    Depends on what data. What I would love to see is data that allows a reasonable prediction rate of job success. That data isn’t coming any time soon though, I feel.

    From the data that does exist, or that I’m aware of to an extent, I’m not sure what it’s good for beyond screening resumes a little more efficiently.

  • http://www.boooleanblackbelt.com Glen Cathey

    I’m sorry but I have to laugh at the idea that analytics (what most people are incorrectly referencing as “big data”) is a fad in HR/talent acquisition.

    There are companies leveraging people/data analytics right now for tangible benefit, both to the employees and the companies, and because of those benefits, the analytics won’t be going away (as per a fad), but advancing and evolving to generate greater benefits.

    I’m not here to convince anyone of anything – I have neither the interest nor the time. Just know that what seems to be a fad for many is quietly being leveraged for a true competitive advantage. I’ve seen it, and it is very real.

  • Richard Araujo

    Glen,

    I don’t think anyone is saying it can’t be, and isn’t already useful. Rather, that there are people who will try to use it to solve problems for which it is not suited, by people who don’t know how to use it or its limits. Like a presumptive carpenter trying to use a powerful nail gun on soft wood, or someone trying to filet a fish with a paring knife. You need the right tool for the right job, and a person who knows how to use it.

  • http://www.ets.org Michael Krause

    Data analytics is not a one size fits all blanket solution and every company can use the information to suit the specific needs. It is not an exact science as yes, there is a human variable to it but it can help tell a story. You can use data to help in resource allocation and to help lay down strategies. Were the 5 conferences we attended last year successful? What percentage of our staff will be retiring in the next 5 years? Do we have an issue with short term turnover? And to an extent I don’t agree that recruiters should be held responsible for turnover however we can use data to see if there are trends. Does one manager or dept. have a significantly higher rate of short term turnover than the company average? If so, there most likely is an issue that needs to be addressed. Data can be used for time to fill purposes as well. And not in the traditional sense of “our time to fill goal is 45 days” and we are judged on hitting that goal. Instead we can take a look at jobs that require more than 45 days. You can use the data you collect to find trends there as well. Are the longer reqs requiring niche skill sets? Does the hiring manager travel and availability is limited? Maybe specs change midway through the recruiting cycle and the original job morphs into something new. Do certain positions see high rates of relocation? If so, why? Once you identify these trends you can work on developing strategies to improve efficiency and effectiveness.

    Big data can also be used in process improvement. You can use data to break down processes and identify weaknesses in those processes. Why is our typical interview process 5 weeks long? Are there redundancies? Can admin work be automated?

    You can’t take the human element out of anything, especially recruiting but data can be an extremely useful tool to identify trends (good and bad), help establish standards, develop strategies and improve processes.

  • Keith Halperin

    @ Richard: I think it would be very difficult to quantitatively predict the success rate of a given individual in a particular job without making a number of major (and probably faulty) assumptions.

    @ Glen, @ Michael: I think it will allow some companies who are currently strategic to improve their results, IFF the presumptions and assumptions under which they make their calculations/analysis are valid and don’t change significantly during the course of prediction period. If you’re not currently able/willing to be strategic, I’m not sure BD will help make it any easier for you.

    Cheers,

    Keith

  • Richard Araujo

    @ Keith,

    I agree, and since that’s the only way it’ll be done what’s needed is a person or company with enough money and the tolerance for the risk/losses that would be inevitable before those assumptions started to align with reality. Which means it’s not likely to happen, so for me, big data is going to be of most value in benchmarking and historical trends, also for real-time market conditions. But for assessing talent I think it’ll be essentially useless.

    That’s what I think this stuff would be good for, letting people know what has happened and what is happening in the strictest objective sense. The why of it all and making sense of it will always be left to human judgement.

  • Keith Halperin

    Thanks, Richard. What I would like big data (“digital dossier” and detailed formal analysis) would be to know as much as about a given person as I could in an easily usable format so as to be better able to recruit them if I wished to. I regard BD as making recruiting more like intelligence work…

    Cheers,

    Keith

  • http://www.steedsgroup.com Gary Steeds

    Well, this article and most of the others as well as these comments about what I still personally call “Big Date HR Techno-Babble” only serve to reinforce my growing fears about this aspect of our business. The only ray of sunshine I have seen is David Bernstein’s article that may show some real direction and possibilities to all these machinations. Let’s hope that wisdom and common sense prevails.

  • Keith Halperin

    @ Gary: “Let’s hope that wisdom and common sense prevails.”
    What do “wisdom and common sense” have to do with recruiting?

    Cheers,

    Keith

  • http://www.equest.com/news/floating-point/ David Bernstein

    Wendell – I wholeheartedly agree that there always has to be a component of HR analysis that is purely focused on the individual. Where we disagree, though, is on the opportunity and the value that Big Data analytics provides to HR in terms of HR’s other responsibility – i.e. being a strategic partner in guiding and running the business.

    Presenting the idea that Big Data analytics is an either/or decision for HR, is a false dichotomy. It’s not about choosing between individualized performance measurement vs. Big Data analytics. Big Data needs to be another tool in the portfolio that HR can tap into to inform the recommendations and decisions the profession is responsible for.

    Big Data enables HR to ask new and different kinds of questions. There are numerous stories in the recent press that point out the insights and successes being achieved through Big Data analysis. I also see this daily in my work with my customers.

    I see the challenge for HR is more of being able to continue to maintain its role as the champion of the individual while at the same time strongly stepping into the role that the Executive Team is needing from HR. How can HR be that “strategic” business partner without all types data and analytics being part of what informs their decisions.

    The HR profession has a long established reputation for being intuition-based in its decision making. Instead of downplaying any particular analytic method, we need to converge on a message that HR needs to embrace a “data everywhere” strategy. That mindset will require the profession to learn how to utilize all data and tools available.

  • http://www.ScientificSelection.com Dr. Wendell Williams, MBA, Ph.D

    @David…Big Data is no different from lusting after a new and catchy statistical program…Garbage-in-Garbage-out… You can think of Big Data in terms of collecting and analyzing sports-team performance. You can plug-in game data all day long, but until you add information about the players, your analysis is limited. When you leave out people, you leave out all your source data…Think that works?

  • Richard Araujo

    @ David

    I think this is where the concern comes up; you need to know what the information means and what you can and can not infer from it. Anyone who doesn’t know that will be just as likely to make a bad decision as a good one with or without any help from big data analytics. In fact I’d say they’re more likely because the “No” vote in hiring is considered the safe path, and people often look for reasons to say no and ignore reasons to say yes. And then, when they do say no, everyone else involved assumes facts not in evidence about the possible subsequent performance of the candidate and assumes a bullet was dodged, when in fact they could have just turned down the superstar candidate of their careers.

    That’s one of the pitfalls with psych tests as well. You present the results and train and help people interpret them, and they still hive off into completely unsupported conclusions. You see this kind of behavior in shopping all the time too, another decision making process. People generally decide what they want based on completely subjective factors and then search out information to justify the decision after it has already been made.

    Data is only as good as the mind analyzing it, and as I’ve said before, the average person is… average. The results they get are likely to be of a similar quality regardless of the tools they are, or are not, using.

  • http://www.equest.com/news/floating-point/ David Bernstein

    @Wendell – I believe we are much in agreement – i.e. the importance of focusing on the individual and the GIGO concept, etc. What I’m also positioning, though, is that the type of analysis that we can now do against large volumes of data from disparate data sources can also reveal insights that are quite powerful/create competitive advantage. It’s not about choosing one approach over another. There are real world examples of employers leveraging this capability today. For example, I know of an employer who reduced OT cost and their attrition rate, and improved their recruiting cycle time by analyzing together tenure, zip codes, and public transportation routes.

    @Richard – I am also a big proponent of your last comments – i.e. the importance of skilled HR professionals to know how to accurately interpret the results of any data analysis – big data or otherwise. I wrote a blog piece on this idea titled, “A Fool With A Tool, Is Still A Fool.”

    Bottom-line – To all on this thread…My view is that the debate is not be about what types of metric or analytics the HR profession should use. I would advocate instead that we will all have more impact on the profession by grappling with what are the best ways for the profession to shift its reputation from being “gut-based” decision makers to being one that is more strategic and evidence based.

  • Robert Dromgoole

    I think the author more than likely didn’t read the book ‘Moneyball’ by Michael Lewis and Billy Beane and the Oakland A’s. Statistics can predict performance, it’s been shown to be true.

    Big Data is just beginning in HR, and it will empower our function like never before.

    Want to know where to build your next factory? Let’s check the talent.

    Want to know how much to pay? Let’s check the data.

    Want to know when to promote? Want to know what your competitors are up to? Which schools are pumping out grads in what space? Who is hiring? Who is leaving?

    Big Data will offer all of this.

    It will require analysis to digest it all, but it will be there at our finger tips.

  • http://www.ScientificSelection.com Dr. Wendell Williams, MBA, Ph.D

    @ Robert…You missed the point of the article. HR generally has little or no trustworthy “human source data” to work from…just results. If you think subjecting bad data to expert analysis will empower your function, then you are in for a rude awakening.

  • Keith Halperin

    @ David, @ Robert: ISTM that you are accepting the premise that HR will be “empowered” or “get a seat at a table” by being less “touchy-feely” and more “number-crunchy”. You don’t get power by counting beans (even huge quantities of them in all sorts of fancy new ways), you get power by playing political hard-ball. If BD will help do that (maybe finding and digging up the skeletons for you to use), then you may have something….

    Keith

    “Power concedes nothing without a demand. It never did and it never will.”
    -Frederick Douglass

  • http://www.equest.com/news/floating-point/ David Bernstein

    @Keith – I’m not suggesting that HR should do less of anything. I do not believe it to be an either or moment. I agree that HR has earned a good deal of political clout. I’m only advocating that HR increase its use and comfort level with all types of data, metrics and analytics; including Big Data analytics.

    Numerous surveys show that Senior HR leader are reporting to their CEO’s. What other business function has the opportunity to touch candidate and employee’s lives, keep the company compliant, operate with low overhead, and be asked to participate in the strategic plans and execution of the company? It’s this last piece that will require an even stronger adoption of data and analytics than what HR generally engages in.

  • Richard Melrose

    @David

    None of the F1000 could pass a routine EEOC compliance audit of their selection processes. Bottom line: they do not uniformly apply valid, job-related selection procedures, backed up by current, properly documented job analyses, within the meaning of the Uniform Guidelines on Employee Selection Procedures (U.S. DoL 1978). That also means that they interview the wrong people, poorly, and then hire somebody.

    The suggestion that HR organizations should “graduate” to something more complicated and less proven, than what their colleagues haven’t been able to master, in over 35 years, boggles the brain.

    As Wendell has noted, previously, most companies waste 20-50% of annual payroll – most of which they can fix, without resorting to big data. As a former international industrial CEO, I say start there; show me the money; demonstrate your competence; then we can talk about Big Data (if you must).

  • Keith Halperin

    @ David: As a recruiter, it is irrelevant to me as to what the Sr. HR leadership does except insofar as it reflects on my colleagues’ and my working environment, compensation, and stability. If BD improves my/our situation, then

  • Keith Halperin

    it’s good.

    -kh

  • http://www.myaspenadvisor.com Andrew Gadomski

    Had to jump in on this one. WOW.

    Big data for HR is not about looking through your HR systems, seeing metrics, and seeing results. Thats little data. Significant, but still little data. The velocity and volume of HR data systems is small. Recruiting may have the most transactions, but learning, performance mgmt, ER, labor, and talent mgmt all produce data at a much slower pace.

    Big data for HR starts AFTER all that data is integrated, and then begins to interface with data produced from the workforce directly (all finance, research, ops, sales, customer service systems), the work product (documents, servers, phone calls, etc) and everything else it can get is hands on from vendors, marketplace analytics, customer behaviors, and so on.

    This is done in the pursuit of finding out if the most expensive resource a company has (human capital) is working the way it could and should. HR has the opportunity to be in the driver’s seat to analyze how well an organization is producing.

    So I pose this – what’s more logical:
    1) tools and knowledge that let an organization know how much productivity, and profitability and waste it has being completely dismissed for its most expensive and long term investment

    OR

    2) all that being placed somewhere other than HR.

    I vote number 2 :)

    Big Data for HR is not going away – but it may be stripped from HR and placed into the hands of others if we don’t step up and get moving on using what we have to make better business decisions.

    - AG