“Unalytics” Awards – Nominee #1

The growing appetite for human capital analytics is drawing more professionals into the wider “market” for human capital analytics, including some players who ought to have done a little more homework.

We hereby announce the Nelson Touch Consulting Awards for Human Capital Unalytics – for the most egregious displays of half-baked notions or products. If analytics represent the use of data, analysis and systematic reasoning to make human capital decisions, then these awards are for the opposite – the misuse of data, poor analysis and fuzzy reasoning to make human capital decisions.  Hence:  “unalytics” – short for un-analytics.

Nominees will be named throughout the year and a 2011 winner will be chosen from among 10 finalists through a reader poll.

We certainly don’t want to embarrass or discourage thought-leaders and vendors who are nominated. After all, it’s the ‘mad’ inventors who eventually end up with game-changing ideas! But at this nascent stage in the evolution of human capital analytics, HR professionals should not be misled by half-baked models that stifle quality and dilute standards.

We welcome defense from the nominees (who may choose to remain anonymous) and debate from the blog’s readership.

Nominee #1 is a human capital analytics software vendor. The company offers customers a dashboard that seeks to warn management of a broad variety of talent management issues, based on analytics calculated using the company’s employee database.  In addition to the analytics, which are presented in graphical format, the dashboard provides an associated color-coded risk assessment.

One of the dashboard items is an exhibit that depicts differences in pay between men and women. On this count, Nominee #1 is to be commended for attempting to throw light on a very important issue. Gender pay disparities have lingered very long without appropriate redress.

The passage of the Lilly Ledbetter Fair Pay Act (2009) renewed interest in the issue, which is probably why pay disparity analysis is showing up in human capital analytics products. Of course, due to the sensitive nature of the topic, analysis is done privately and there is little scope for benchmarking.

However, Nominee #1 doesn’t give us a practical or accurate approach to investigate gender pay disparities. There is no happy intersection between the surge in human capital analytics (the tool) and the addressing of gender wage dynamics (the issue). Instead, here is Nominee #1’s ham-handed solution, depicted below.

The graphic is confusing and, stunningly, at once both overly complicated and overly-simplified.

It is overly complicated because what’s important here is the differential between male and female pay. This can easily be shown as one number: average female pay as a proportion of average male pay. This happens to be 60% in this example. In one number, 60%, you immediately see the overall problem (it’s not close to 100%) and the extent of it (it’s not even close). No need for a chart at all!

The graphic is overly simplified because comparing gross averages between men’s and women’s pay does not provide us much useful information. Certainly, knowing the gross extent of the gap is a start. However, the gap needs to be decomposed into what is explainable and what is not. Differences in labor market experience, educational qualifications and specialized skills (i.e., individuals’ human capital stock) might account for some of the gap.  The unexplainable portion of the gap is the problem and most of it is attributed to gender wage discrimination.

The actual problem of unjustified pay disparity might be different from what one might assume looking at just the gross difference. Furthermore, there might be important insights to be drawn from the more detailed analysis. It could be that men and women earn different rates of return on their individual human capital (which needs further exploration). The main point here is that it is misleading to try and boil down this important issue into one simplistic dashboard statistic.

And then there are some minor irritants in terms of the graphical representation.  Why are we burdened with two decimal places of significance when the gross difference between the categories is so large? Reporting no decimals or at most one decimal place would be sufficient. Why is the pay differential denominated in 22% increments on the y-axis? Not adjusting what seems to be an automatic scaling setting shows a disregard for the numbers and our analytic sensibilities. What elements are included in “Pay”? Is this for a specific position or is it an average over all positions? Why are we not provided trend information on a statistic that we want to improve over time?

Finally, we come to the risk assessment, represented alongside the graphic as follows.

The differential represented here is that between average female pay and overall average pay, when the crux of the matter is the differential between average female pay and average male pay. One explanation for this approach might be that builders of dashboards need to provide a baseline or target and, in this case, overall average pay appears to have fit the bill.

The differential, again to an unnecessary two decimal places of significance, is flagged as a “Severe Risk.” Once more, I think the dashboard construct has trumped thoughtfulness about the issue, the metric and the risks posed.

Have I missed anything? Please join the discussion and stay tuned for Nominee #2 for the Nelson Touch Consulting Awards for Human Capital Unalytics. You are invited to submit “unalytic” gems that you come across for deconstruction and debate.

 

Posted in Analytics, Compensation, Incentives | Tagged , , | 1 Comment

Leadership vs. Management

Are these two notions distinct, synonymous or complementary? Many views prevail in the literature.

I find the views of John Kotter and Peter Northouse particularly compelling, based on my own experience with leaders and managers at all levels of organizations, especially during times of change.

Kotter argues that leadership and management involve two distinct but complementary sets of action. Leadership is about coping with change while management is about coping with complexity. Here is a summary* that I keep handy to distinguish between the two.

* From Peter G. Northouse’s Leadership: Theory and Practice, Fourth Edition (2007) in which he draws from John Kotter’s A Force for Change: How Leadership Differs from Management, (1990).

 

Posted in Change Management, Leadership, Organization Design | Tagged , , , , | 3 Comments

Text Analyze This

What are the different connotations ascribed to the term “human capital” by HR professionals and what is a simple way to illustrate the variety?

The opportunity to examine these two questions came up recently as part of some work with Society for Human Resource Management’s (SHRM) Measures and Metrics task force.

The Measures and Metrics task force is one of many that have been convened to establish standards around HR metrics. This innovation over past attempts is the employment of the American National Standards Institute (ANSI) protocols for standards development.

The Measures and Metrics task force is drawn from HR practitioners, consultants and other interested parties from around the world. Its remit is to develop measures and metrics that will be useful for investors – ones that might become standard elements of the United States’ Securities and Exchange Commission’s (SEC) Form 10K, for example.

The term “human capital” is ubiquitous, but it often means different things to different people. The term was invented by economists (see our previous blog post, “A Capital Idea?”), but has now entered the business lexicon and is often championed by HR.  It was a natural choice as the basis for the task force’s metrics nomenclature

In order to ensure that everyone was talking about the same thing, members of the task force were asked to write down their views on what they understood by the term “human capital.” Many members responded.  Reading through the definitions, it became clear that there were disparate views on what human capital meant or ought to mean.  As to be expected with such a popular term, there were some common themes and words – education and experience, for example – but also some new words and ideas.

I thought it would be a good idea to create a “word cloud” (similar to the “category cloud” on the right margin of this blog) of the submissions. I used a free web-based application called Wordle, though there are many such free text analytic applications available.

The image you see on the left of the page shows the words used in the submissions to describe “human capital.” The size of the word is proportional to the number of times the word is used in the document. This particular application shaped the words around a footprint shape (this is a customizable feature and you can choose from a variety of shapes and styles).

Right away, you can tell which are the most common words used. Such text analytics can be very useful. From this blog’s tag cloud, for example, you can see which are the most popular topics discussed.

There are a number of other possible uses.  HR leaders can look at HR communications to ensure that the message is not being overwhelmed by certain words and phrases. Compensation professionals can examine job descriptions to check for an appropriate balance in the verbiage between strategic and operational responsibilities. Team leaders can ascertain the common themes among team members’ inputs on various topics. Job seekers can ensure that their resumes are hitting the right notes in terms of key words and capabilities.

Text analytics is a growing field and has come a long way. It will grow in importance as the web evolves towards a semantic structure and capability (the “semantic web”).  Words in a document are now akin to numbers in a spreadsheet. Evidence-based HR enthusiasts should familiarize themselves with text analytics and leverage them as they would quantitative analytics – to help answer critical questions and make better decisions.

 

Posted in Uncategorized | Leave a comment

It’s Probable: You Have a Chance

As HR professionals wrap their heads around predictive human capital analytics, one of the capabilities required will be a firm grasp of probability. Not just relevant probability theory, but a feel for the numbers as well.

Why is this important? The outputs of predictive human capital models are typically probabilities, likelihoods and odds. If we are going to use predictive models, we need to wrap our heads around these special numbers.  This post deals with probabilities. A probability is simply a way of expressing knowledge or a belief that an event will occur or has occurred.  Subsequent posts will cover likelihoods and odds to complete the discussion.

In most cases, we will have an inkling or some sort of gut feel for what the prediction of a predictive human capital analytical model will be, based on training and experience. For example, if we are trying to model the probability of a new hire’s success in the company, based on his or her attributes, we typically have a feel for what sorts of candidates succeed in different business units, based on observations over the years.  Our predictive model will quantify this probability so that we can judge the impact of various factors.

Unless we are familiar with the mathematics underpinning the model, however, we might be surprised by the output. Unexpected results should always be examined to unearth the source of discrepancy from our expectations.  Sometimes the problem is the model; sometimes it is our expectations which are awry.

Two examples have emerged in the popular literature to illustrate how we cannot always trust our initial instincts with regard to probabilities.

The Birthday Problem

The first example is known as “the birthday problem.” It has been around for a long time, but I was first confronted with it in an applied mathematics course at graduate school. It was one of the problems in the problem set assigned at the very first class. Fortunately, the solution was discussed by the professor in the next class (and fortunately for current students, is now widely available on the internet).

The problem is to figure out what is the probability that two people in a group of random individuals have the same birthday (day and month). Obviously, the larger the group, the larger is the probability.

Most people underestimate the likelihood that two people will have the same birthday. We have a prior notion that birthdays are rare (you have to wait 364 long days until your next one) and we only seldom encounter people with the same birthday as ours.

If you do the math right (there are a number of ways to arrive at the result), you get some fairly counter-intuitive results. With 20 people in a room, the probability of a shared birthday is as high as 41%. Increase the number to just 23 and you have even odds of a shared birthday (50% probability).

Looking at it another way, you only need 57 people to reach a 99% probability of a shared birthday. The table below summarizes the relationship.

Try it out at your next meeting. If you don’t get a shared birthday, remember this was a probability – you are not guaranteed the result. Even with 366 people in the room, if one person has a birthday on February 29, all bets are off.

The Drug Test Problem

The second example is typically framed as a drug test situation. Imagine a drug test that is known to correctly identify a drug user as testing positive 99% of the time and correctly identify a non-user as testing negative 99% of the time. That’s quite accurate by any measure. Now let’s assume that 0.5% of employees are drug users.

When the math is done right (using Bayes’ Theorem), the probability that someone who tests positive for the drug is actually a drug user is – hold your breath – 33%!  It’s more likely that the person is not a drug user!! Surely this can’t be – what’s going on here?

Despite the apparent sensitivity of the test, the low rate of drug use results in low accuracy. Basically, there’s a greater chance of false positives when the use rate is so low.

Both the birthday example and the drug test example are covered in greater detail in Wikipedia, including all the underlying math.

What are the conclusions? Here are two.

First, you need to have some understanding of basic probability theory. Being able to compute an expected value is essential. Knowing the difference between a probability and a conditional probability is important. One easy example is estimating turnover. Most people would measure turnover in a period as a percentage (say 17% if 17 people out of 100 left the organization). They would perhaps estimate that the probability of turnover is 17%. Fair enough. However to get an accurate estimate, one needs to look at individuals’ tenure and compute the probability of turnover after x years. This probability needs to be conditioned on the fact that they have not terminated for x years. I’ll cover this “life expectancy” notion of turnover in a subsequent post.

Second, gut feels for probabilities only work sometimes. It would be advisable to do the math or compute the model and come up with the actual probability. Of course, if the number does not agree with your hunch, you need to be able to get under the hood and work out the probabilities formally in order to convince yourself which number is right. With multiple variables and complicated or “advanced” predictive models, it is going to be very difficult to do the math yourself. At some point you have to trust the model. Which only means that you need to understand how he model works; it’s not enough to press a button, get the output and stick it onto a Powerpoint chart. Odds are that someone will call you on the probability underlying the prediction.

 

Posted in Analytics, Compensation, Incentives | Tagged , , | Leave a comment

A Comment on Pay Transparency

The following is a letter to the editor of HR Executive Magazine published in the Jan/Feb 2011 issue.

As the article suggests, pay transparency is more than just revealing everyone’s pay. Rather than opening a Pandora’s box of emotions and unhealthy thoughts (“Ha – I’m way better than X” or “Waah – I can’t believe the injustice of it all!”) that will consume managers’ time and energy, pay transparency should be focused on everyone understanding the rules of the game.

When employees have the context of how salary and bonus budgets are determined, how individual pay outcomes are decided and how those outcomes relate to their performance relative to their peers, there is likely to be much less angst and awkwardness about pay discussions and employees can work on improving their future pay outcomes.

However, a lot of groundwork needs to be in place to ensure that all of this actually works. The company culture needs to emphasize openness and trust, employees need to have faith in the fairness of goal-setting and performance evaluation and there must be a proven history of pay-for-performance compensation decisions.

Posted in Compensation | Leave a comment

Size Matters

In his Research for the Real World column in World at Work’s Workspan magazine, Kevin Hallock asserts that CEO pay is positively correlated with firm size, even when controlling for a number of other variables. Should we be surprised? I don’t think so.

Hallock’s assertion is based on research he did for The Conference Board using CEO pay data for 2,300 US companies. Total compensation (salary, bonus, equity) ranged from $878K in the lowest revenue decile to $10.2 million in the highest revenue decile.  The relationship holds even when other variables are introduced into the analysis – firm-level variables such as industry, profitability, R&D expense and individual-level variables such as age and experience.

Hallock introduces the notion of elasticity – the percentage change in one variable associated with a percentage change in another – and estimates the elasticity of CEO pay with respect to firm revenue to be 0.3. The elasticity is ostensibly the coefficient on the revenue variable in a regression that includes the control variables mentioned above.

The quantification of the relationship (for this specific data set) is valuable. However, we should not be surprised that CEO pay increases with firm size.  Why not?

Compensation is based largely on job scope and performance. In the case of CEOs, a larger organization is essentially larger job scope. If a firm acquires another firm, the acquiring firm’s CEO’s pay will either stay the same or increase (maybe not immediately, but you can be sure that the case will be made and viewed favorably by the compensation committee). The pay premium for increased size reflects the increased responsibility and risk that the CEO now bears.

In practical terms, compensation levels are determined through salary surveys. Most salary surveys segment participation by firm size, recognizing that pay level vary across size categories. It’s a little bit of “chicken and egg” – did the surveys simply capture the extant firm size vs. pay relationship (presumably the first ever survey did) or did survey data subsequently drive the relationship?

Are there any avenues for research that might throw additional light onto the correlation of firm size with pay? Here are a few suggestions.

  1. Hallock’s results are for CEO’s in US companies, so a natural question is whether these results hold outside the US and if so whether the elasticity of CEO pay with respect to firm size varies by geography.  What specific factors drive the variation?
  2. It’s likely that the relationship holds for other executives, but what about non-executive roles? Surely there are some benchmark roles whose pay is not as strongly correlated to size as is executives’ pay. One might surmise that the HR function’s pay does not increase as dramatically as say the risk management function’s pay across investment banks.
  3. Does it matter how you measure firm size? Hallock’s work is based on revenue. Do other measures of size such as assets, market capitalization or number of employees exhibit a stronger relationship? Hallock claims that the relationship is robust to the choice of size metric. Perhaps one metric is a better predictor of pay? Survey vendors take note.
  4. Does the macro relationship (i.e., at the firm level) hold at the micro level (i.e., divisions and departments within a company)? It ought to, if our rationale above (job scope) is correct.
  5. How does the elasticity of pay with respect to firm size vary with firm size?  I suspect that Hallock reports elasticities at the mean; elasticities are different at different levels of firm size.
  6. Has the relationship between firm size and pay evolved over time?

There remains plenty of work to be done to fully characterize the relationship between firm size and compensation. Why is this important? We’re always seeking to understandd what drives pay and therefore be in a position to predict what the “right” pay level should be.

Survey vendors that have run compensation surveys across multiple industries for many years are in a position to examine some of these questions. It’s a pity that they tend not to capitalize on the vast treasure trove of information they possess. I’m sure their customers would be interested in some cross-company analytics such as elasticities of pay with respect to firm or division size.

World at Work is to be commended for including a research-oriented column in Workspan and Hallock is to commended for introducing interesting ideas.

Posted in Compensation | Tagged , , , | Leave a comment

Deeper into Box Plots

In the Behold the Box Plot post, we examined a simple box plot, which is good for most purposes. The version that most statisticians use, which I’ll call the statistical box plot has two additional features.

The first, is to show the average or mean of the distribution in the box. The second is to characterize some observations as outliers and show them explicitly. Figure 1 illustrates the statistical box plot. I will explain each part of this new schematic.

The inter-quartile range (IQR) is the distance between the upper and lower quartiles and contains (by definition) 50% of the data points around to the median. The average, denoted by a dashed line, completes the picture of the data distribution.  Compensation professionals rely more on the median, since it is a measure of central tendency that is not influenced by outliers.

It is important to identify data points that are far from the median. These may or may not be relevant to the analysis.  When looking at salaries, for example, a person with an extremely high or low salary may be a special case and therefore outside the scope of the analysis. In general, the focus of any analysis is where the bulk of the data are – those data points within the IQR.

The IQR is used as a metric to denote two ranges of data that are illustrated in Figure 2.  The “inner fence” is 1.5 times the IQR beyond the lower and upper quartiles (i.e., beyond the box).  The “outer fence” extends 1.5 times the IQR beyond the inner fence (i.e., 3 times the IQR beyond the box).

These ranges allow us to classify data points beyond the IQR as “outliers” or “extremes.”  An outlier is any data point that is beyond the inner fence but within the outer fence. Outliers are denoted by an “x” symbol. An extreme is any data point that lies beyond the outer fence. Extreme values are denoted by an “o” symbol.

As you can see in Figure 3, the lines that extended from the box no longer reach out to the minimum and maximum of the data. These lines are also known as “whiskers” (box plots are sometimes referred to as box-and-whisker diagrams).

In the statistical box plot, the lines extend to next data point outside the box. If there is no data point within the inner fence, then the line extends to the inner fence (there can still be outlier or extreme data points).

 

There is a lot of terminology involved here, but the picture itself speaks a thousand words. Once you are familiar with the elements of a statistical box plot, you can glean all the relevant features of the underlying data distribution at a glance.

Statistical packages such as Stata, SAS and SPSS allow you to customize the box plots to various degrees.  Each uses slightly different terminology for each of the elements we have reviewed so far. However, now that you have the general idea, you should be able to navigate the peculiarities of any software that generates customized box plots.

One useful feature is to vary the width of the box by the data size (i.e., number of observations) so that you can visually weight the data when comparing different distributions. If you are not interested in the outliers and extremes, you can suppress them from the graphic.

Box plots can be used to characterize data distributions in any field. Figure 4 shows the results of the Michelson-Morley experiments of 1887 to measure the speed of light.  Of course, the box plots were created a century later, having been invented in the late 1970’s.

There is an even more elaborate way to represent a distribution (hint: it is named after a musical instrument). I’ll defer that description until we’ve taken a look at probability densities and cumulative distributions. Stay tuned!

 

Posted in Analytics, Compensation | Tagged , , , , , , , , | Leave a comment

Book Review: HR Analytics Handbook

The field of human resources analytics (aka human capital analytics) has taken the HR world by storm.  The unsung analytical heroics of rare HR quants, hitherto hidden from sight, have moved to center stage.  HR professionals worldwide have embraced the use of data, analysis and systematic reasoning to make human capital decisions.  The literature, blogosphere and conference circuits are rife with articles, books, posts, twitterings, Linked In groups, webinars, courses and conferences related to human capital analytics. What is to be made of all of this?

“HR Analytics Handbook” by Laurie Bassi, Rob Carpenter and Dan McMurrer arrives at the right time and with the right focus. It is targeted at the reader who recognizes that something major is afoot in the practice of HR and wants to get up to speed on the topic. Bassi provides a clear, concise and practical briefing on the state of knowledge in the world of human capital analytics. Readers will be well rewarded for their investment of time in studying the handbook.

The handbook prompts the reader to consider his or her next step with respect to HR analytics:  stay away, stay abreast, get involved, or take the personal initiative in terms of projects or career direction.

The book is a quick read at 54 hand-sized pages (excluding endnotes and bibliography). The five sections cover the definition of HR analytics; how to get started, needed skills, and common pitfalls; recent empirical findings; examples of organizations using HR analytics; and conclusions.

The material is well organized and the presentation is succinct. Since Bassi and her co-authors readily admit to modest intentions with this brief primer and accomplish their objectives,  one cannot find fault with their endeavor.

However, as a literature review the handbook is not exhaustive even though the universe of publications is fairly compact at this point in time. It does not consider the vast trove of often insightful material (albeit largely commentary rather than actual analytics) in the blogosphere. The additional readings listed at the end of the book are thoughtful, but could have been arranged by topic to guide the lay reader on the appropriate sequence of further study.

The handbook does not describe the toolkit available to human capital analysts beyond mentioning correlations and regressions in passing. It would be helpful to know the specific statistical techniques and protocols that can be applied, as well as to distinguish between “model building” and “data mining” as alternative analytical approaches. Models are based on a notion or theory about how variables inter-act which are then tested statistically using relevant data. Data mining is theory-agnostic and is based on discovering patterns in the data.

One simple yet powerful notion underpinning the application of analytics to human capital decisions – hypothesis testing –  is not mentioned.  Analytics allow us to formulate opinions as hypotheses and to design tests based on data and statistics that can determine their validity. Hypothesis testing is an elemental form of analytics that leads to better decisions and, sometimes, results in spectacular myth-busting.

All the above notwithstanding, Bassi and her co-authors are to be commended for taking the first step in reporting on the state of human capital analytics. Their timely contribution challenges their peers to fill in the remaining blank spaces in the literature.

There is a need for an introduction to the actual practice of HR analytics targeted at HR professionals that guides them through the tools and methodologies applied to relevant areas.  Yes, it means getting more comfortable with data and statistics and perhaps reaching out beyond Excel to some specialized software. I’m happy to report that yours truly has undertaken this book project.

 

Posted in Analytics | Tagged , , , , | 1 Comment

Behold the Box Plot

Compensation professionals are constantly examining the distribution of data – e.g., the distribution of salaries, salary increases, bonus payouts, etc. While typically focused on quartiles and medians of the distribution, they recognize that they need to look at the entire distribution to truly understand the data in question and recommend the right decisions.

The box plot is one of the most widely used devices for examining and comparing data distributions. Due to their almost exclusive reliance on Excel for data analysis, however, the vast majority of compensation professionals have not been exposed box plots (Excel does not offer box plots in its graph portfolio) and their usefulness. This post attempts to persuade compensation professionals and other human capital analysts to consider introducing box plots into their analyses and presentations.

The important features of any data distribution are the variation of data and their clustering, if any, around certain values. The variation is characterized by the range of data as determined by the minimum and maximum values.  If the clustering typically occurs around one value, this is the central tendency of the data. The other important feature is the spread – the degree to which the data cluster around the central tendency. If there is only a little clustering, the data have a large spread.

For small data sets, the astute analyst can typically get a sense of the distribution by looking at the column of numbers (sorting them in ascending or descending order helps a lot). However, for large data sets, you would need to plot a histogram. Even then, some more work would be required to identify the quartiles. This is where the box plot comes in – it portrays the data distribution in a simple graphic that displays all its important features. Figure 1 shows a simplified box plot.

The box represents the 50% of the data between the 25th and 75th percentiles (i.e., the lower and upper quartiles) and draws focus to the center of the distribution, the median, which is depicted by the red line inside the box. The vertical lines extending from the box reach out to the minimum and maximum values. Looking at two different distributions via their box plots as in Figure 2, we can compare their central tendencies, spreads and ranges at one glance.

The box plots are much more striking and informative than the tabular data.

How might box plots be used in compensation? Figure 3 illustrates how internal salaries are arrayed in relation to salary ranges and the market data. Again, in one quick glance you can see where the outliers are (both internal and versus the market) and how the internal salaries compare with the market salaries. It is unfortunate that most survey vendors do not illustrate market data using box plots and are not in a position to report back participants’ overall salary distribution in relation to the market.

I referred to the box plots depicted above as simplified box plots. A subsequent post will explain variations on the simplified box plot.  I should mention that the box plot is a fairly recent invention. It was introduced by John Tukey in 1975 and has been a staple of statisticians since then. Box plots are used widely in medical research and economics and are a standard graphing format in statistical software programs.

In a subsequent post, we’ll examine how the average of the distribution and outliers are depicted.

 

Posted in Analytics, Compensation | Tagged , , , , , | 1 Comment

Value Proposition for an Early Investment in Human Capital Analytics Capability

A few weeks ago I was visiting with an HR executive at a small but growing technology company who asked me about the value proposition for investing in a human capital analytics capability. The company uses analytics in its core business and the leadership team is inclined toward human capital analytics. In fact, it was ready to fund the development of such a capability within the company. My host was seeking an opinion on how a human capital analytics group could justify its existence so early in the company’s life. Here is what I told him, more or less.

The immediate returns to analytics are higher for larger companies (more data, larger dollar consequences of bad decisions, etc.) so unless there is a burning platform requiring instant attention, thoughtfulness at the outset about the size and nature of the investment is smart. Too many companies are jumping onto the human capital analytics bandwagon without adequate due diligence. “It’s the latest thing…everyone else is doing it…HR needs to be more strategic…we can be like Finance” – these are not useful considerations. Deciding to be an analytical function is a strategic move, not a tactical one.

With just a few thousand employees and a few years of history, there may not be a lot of data to generate compelling analytics initially, but given the expected growth trajectory, human capital analytics will become a critical management tool. Establishing a beach-head early in the game will catalyze the establishment of a human capital analytics infrastructure. Leveraging insights and support from the business analytics group will be important. There may be some healthy exchange of ideas, data, models and even people in the future.

Getting the systems plumbing into shape early on will pay large dividends later.  Pinning down data sources and integrating them across the enterprise for ready access by analysts can frustrate analytics initiatives in larger companies. It may be prudent to use the same analytical software or systems and database architectures as the business analytics group.  Advantages include price economies of scale on software licenses and lower barriers to collaboration between the human capital and business analytics groups.

Another aspect of establishing the infrastructure in terms of the data themselves is an emphasis on data collection – the importance of actually collecting data and ensuring their quality. The power of human capital analytics is often constrained by the lack of important individual-level human capital information such as education, certifications, previous employment history, etc. Data considerations encompass information on candidates, employees and terminations.  Adjusting processes to generate relevant data and systems to accommodate the data would be important first steps.

Employee perspectives are an important dimension of human capital analytics. They provide a barometer of employee sentiment, engagement and alignment. Employee sensing helps to prioritize programs, capitalize on employee preferences and even predict responses to environmental or programmatic changes. A nascent human capital analytics function can begin to design appropriate employee surveys and other employee-sensing approaches. The longer the time-series data on employee perspectives, the more useful are the overall data. Too many companies don’t ask the right questions or ask the right questions in a way that proscribes the analysis. A thoughtfully designed employee survey approach yields richer and more useful information.

Starting a human capital analytics function early in a company’s life-cycle allows for the function to evolve according to the needs of the organization. Initially, a centralized function can build the infrastructure and set standards for data, analyses and reporting. Once the company’s divisions and geographies are large enough to need and sustain their own analytical functions, the analytics capabilities can be decentralized. The analysts are closer to the people and business and are able to provide customized immediate and local decision-making support.

Introducing the culture of fact-based analysis and evidence-based decision making within HR will lay the foundation for a sophisticated and savvy HR function that is attuned to the way business decisions are made. Analytic capability can be built up as a core competency within HR. It will be important, however, for HR not to succumb to extreme reliance on numbers and technical analysis. It will then be indistinguishable from Finance! Too often in the rush toward human capital analytics adoption, HR’s consummate skills in working with individuals and teams and managing all the “soft squishy stuff” (culture, motivation, performance, etc.) is overlooked and undervalued.

It is worthy of note that the company’s leadership looked to HR – and more specifically to the compensation and benefits function – to develop the human capital analytics capability rather than the established business analytics group (here’s why that’s a smart idea). In most companies, HR might not even be considered for this leadership role since analytical capabilities, especially at the level required for serious analytical work, is seldom available in HR (that’s just one of our problems). However, there are some rare finds in the labor market that combine analytical abilities with HR experience.

Last but not least, to maintain leadership’s buy-in and support of the human capital analytics function, it will be important to help make decisions and solve problems rather than just provide reports. The initial focus should be on tackling important issues through innovative data collection, smart analysis and sober (i.e., not “Look at these fantastic β’s!!!”) presentation. Beware an initial focus on dashboards.  Regular, standard reports tend to lose their gloss after the first few installments and tend to be filed away very quickly.  The sure way to make the business leaders feel they have made the right decision with regard to their early analytics investment (and to ensure continued and even increased funding!) is to have deliver a sound and high-impact business result.

 

Posted in Analytics, Leadership, Organization Design, Strategy | Tagged , , , , | Leave a comment