Machine Learning Times
Machine Learning Times
Video – Alexa On The Edge – A Case Study in Customer-Obsessed Research from Susanj of Amazon
 Event: Machine Learning Week 2021 Keynote: Alexa On The Edge...
Why AI Isn’t Going to Replace Data Scientists Any Time Soon
 Should data scientists consider AI a threat to their...
“Doing AI” Is a Mistake that Detracts from Real Problem-Solving
  A note from Executive Editor Eric Siegel: Richard...
Getting the Green Light for a Machine Learning Project
  This article is based on the transcript of...

5 months ago
Wise Practitioner – Predictive Analytics Interview Series: Haig Nalbantian at Mercer


By: Eric Siegel, Founder, Predictive Analytics World for Business

In anticipation of his upcoming presentation at Predictive Analytics World for Business Livestream, May 24-28, 2021, we asked Haig Nalbantian, Senior Partner, Co-leader Mercer Workforce Sciences Institute at Mercer, a few questions about their deployment of predictive analytics. Catch a glimpse of his presentation, Value versus Performance: How Advanced Analytics Can Help Distinguish Between them and Why it is Important, and see what’s in store at the PAW Business conference.

Q: In your work with predictive analytics, what behavior or outcome do your models predict?

A: Since 1994, our Workforce Sciences team has been applying advanced analytics to help identify and measure the workforce drivers of critical workforce and business outcomes. The goal is to help uncover the sources of human capital value in organizations and use these learnings to inform decisions about workforce strategy and management.

Typically, the core workforce outcomes we model are retention, promotion, performance and pay; sometimes we model employee engagement as well. The business outcomes we model are workforce productivity, operational productivity, profitability, revenue and revenue growth, and measures of customer value such as customer retention, satisfaction and local market share. When we focus on business outcomes, our aim is to estimate how variations in workforce characteristics, outcomes and practices drive changes in business performance and draw practical implications for optimizing workforce strategies.

My focus in this interview and conference talk is on individual and firm performance. I want to illustrate how we use advanced analytics to adjust conventional “raw” measures of performance, whether at the individual, group or organizational levels, to remove the contamination of “situational” factors or pure randomness that undermine their utility as indicators of actual contribution and employee value. Properly deployed as the basis of rewards, these adjusted measures can strengthen performance incentives and better allocate performance “risk” – that is, the sources of variation in performance that reflect the operation of exogenous factors and pure randomness. Perhaps most importantly, adjusted measures can help improve talent assessments by identifying those who perform well given the “hand their dealt” and those who may look like good performers based on raw performance measures but, in effect, are simply lucky. The result, ultimately, is to increase shareholder value because the organization is better able to secure its best talent and motive them to perform well.

My core message is that with the right data and methods, organizations can do a much better job of teasing out true performance from the effects of circumstance.

Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?

A: We work with client organizations – typically larger, global firms – to embed the principles and methods of evidence-based workforce management in their decision-making processes. We have applied this approach to many areas of workforce strategy and management. One area where the approach has proved particularly effective is with respect to Diversity, Equity and Inclusion (DEI). In our experience, a major obstacle to the success of DEI strategies is the prevalence and persistence of significant and sometimes sizeable disparities in performance ratings received by different demographic groups – particularly between people of color and whites. Since ratings commonly influence promotion, retention and pay as well, such disparities have cascading effects that make it especially difficult for diverse talent to thrive at work.

This pernicious pattern remains the norm despite increased corporate commitments to diversity and recognition of inequity at work as a business and social scourge. By identifying and quantifying systemic sources of disparity, we can help organizations address the challenges at the source and create a blueprint for leaders and managers to change realities on the ground.

Q: Can you describe a quantitative result, such as the predictive lift of your model or the ROI of an analytics initiative?

A: A recurrent finding is that an employee’s performance rating may have less to do with their individual attributes and behaviors – e.g., experience, education, background, diligence – than with where they sit in the organization – e.g., the part of the business with which they’re affiliated, the role they are in, whom they report to, the characteristics of their team, among others. Statistically, this means that the same individual placed in a different situation would see a marked change in their raw performance and how it is appraised.

This can be a profound insight for executives as they tend to be extremely performance driven. How often have I heard chief executives expound on being “results oriented” and wanting that mindset to permeate their business. As a result, they are inclined to embrace a pay-for-performance approach to rewards. Not many would argue with that proposition on its face. It seems aligned with the idea of requiring accountability from employees. If only it were that easy.

When confronted with the reality that raw performance measures, including performance ratings, are too strongly influenced by circumstantial factors and noise to serve as reliable indicators of actual performance and value, the lights go on and sometimes these executives start to see their world differently. Taking a more expansive view of performance and using the results of good analytics to develop more precise and informative measures of performance can be transformational in its impact on the business.

To me, this is what workforce analytics delivers at its best. Analytical results can change how leaders and employees see their world, how they think and how they behave. Speaking personally, I can tell you those special moments of sudden illumination with executive teams are among the most gratifying and fulfilling moments in my own life as a consultant. It’s what I live for professionally, especially when the resulting actions lead to better outcomes for the business and its employees.

Q: What surprising discovery or insight have you unearthed in your data?

A: Continuing with my focus on the subject of distinguishing employee contribution and value from raw performance, there are two important insights I want to emphasize here:

The analytical work often reveals how ineffective and costly the typical “crude” pay-for-performance reward systems can be. These negative outcomes show up in higher turnover, lower engagement and, yes, ultimately higher compensation cost. Costs rise because when pay becomes more volatile due to performance risk, employees bearing such risk will require higher expected compensation. Simply put, people don’t bear non-diversifiable risk for free. Competitive labor markets force adjustments in pay to compensate for such risk. That’s an established economic regularity, but one that is often lost on compensation professionals.

Second, recognizing superior talent whose actual contribution is camouflaged by the raw numbers is eye opening, even exhilarating. Imagine what it feels like to find employees who manage to outperform statistically-determined expectation despite being dealt a bad hand at work? Is there a better indication of value? These are the unseen diamonds in the workforce that can propel business growth and profitability. Too often, they fly under the radar or, worse, are devalued and driven to leave.

For the Baseball fans among your readers, it’s equivalent to suddenly appreciating the superiority of the sabermetric measure, “Wins above Replacement” (“W.A.R”) over the traditional performance measure, “Wins” or “Winning Percentage,” as a measure of a pitcher’s value. Whereas “Wins” reflect more about the impact of the team and environment – e.g., the team’s offensive production, defensive skills, the dimensions of the home field, etc., “W.A.R” hones in on what the pitcher has actually accomplished in face of those surrounding realities. Apples-to-apples comparisons make it a more informative measure of true value. Once you’ve moved from “Wins” to “WAR,” there’s no going back.

Let’s face it, most pay-for-performance systems in business are still relying on the equivalents of “Wins.” In my view, that needs to change; otherwise, most organizations would be better off without pay for performance at all.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: I will share some insight about the intersection of good performance measurement and management with efforts to strengthen workforce diversity. There is another stark reality we see in our work: all too often the situational factors that increase the likelihood of receiving high performance ratings and which accelerate career advancement are skewed demographically. In a word, Woman and people of color are not being positioned to succeed in organizations in the same way as their male and white counterparts. I don’t want to suggest this is a universal finding, but I see it too often in the analytical results not to raise the red flag.

For clients, this too can be illuminating. It opens the door to developing a blueprint to eliminate systematic differences that impede the realization of diversity and equity objectives. Here again, good analytics play an essential role, uncovering the facts about what most influences performance and how those factors can be managed to eliminate disparities and render more accurate assessments of talent. Together, effective and unbiased talent assessment and performance management become the foundation of a successful DEI strategy.


Don’t miss Haig’s presentation, Value versus Performance: How Advanced Analytics Can Help Distinguish Between them and Why it is Important at PAW Business on Friday, May 28, 2021 from 10:20 AM to 11:05 AM. Click here to register for attendance.

By: Eric Siegel, Founder, Predictive Analytics World for Business

Leave a Reply