Haig R. Nalbantian, the dean of Workforce Science, chats about his team’s approach to workforce analysis, technology displacing human creativity, the volume of task-oriented activities, the impact of the pandemic and remote work.
Haig R. Nalbantian, Senior Partner and co-founder/co-leader of Mercer Workforce Sciences Institute bridges the world of data analytics and creative forecasting as companies plan for the future. Haig has been instrumental in developing a “new science” of human capital management — “Workforce Sciences.” He and the Mercer team analyze workforce and business performance data to measure and identify where businesses derive value from their human capital and determine which management practices effectively increase this value.
Prometheus: At Mercer, you have an interdisciplinary/multidisciplinary strategy & analytics team focused on developing and implementing methods to measure the workforce and business impact of people practices. How does that team operate?
Haig: We launched this team in 1994, first as an R&D unit, then as a full-fledged business practice. We’re roughly half and half economists and organizational psychologists with a sprinkling of data scientists, statisticians, and a few other eclectic disciplines (e.g., Biostatistics, aeronautical engineering, political economy) in the mix. We’re all about complementarities in skills and experience. We’ll examine modeling results from case data applying the interpretative lens of our different disciplines.
The economists might put forth a possible explanation of a particular finding or broader patterns observed in the data, and the psychologist says, well, you know what, here are other possible explanations based on models from our discipline. So, we’ll have competing ideas to explain and interpret the same empirical pattern. That’s key to our being successful. We have a process of interactive analysis where alternative hypotheses based on both disciplinary knowledge and client observation are developed, deliberated and, where possible, empirically tested before they are synthesized into a meaningful story line that enables executive decisions.
Let me provide you an example: In a large regional bank, we found that a primary driver of employee turnover was the departure of the employee’s supervisor. Specifically, if an employee’s supervisor left it doubled the probability that the employee would leave in the next year. Now that is a statistical relationship, but what’s behind it?
The psychologists on our project team offered that this might be about the disruption of personal relationships at work that provokes reflection on career change. The economists invoked the “signaling model” from economics. The idea is that the departure of a supervisor from the group conveys information about the relative value of career opportunities within the firm compared to opportunities in the outside market.
To evaluate these alternatives, we fashioned a further test to distinguish what happens when the supervisor leaves the employee but does not leave the firm. We found that the cascading impact on supervisor turnover only manifests if the supervisor left the bank; there were no effects at all associated with an internal transfer. As such, we had cause to embrace the signaling interpretation. Apparently, the disruption of personal working relationship were not decisive in this organization. It was all about information signaling regarding market opportunities. In another organization, we might find the opposite.
And that’s the point of this approach. Each organization has unique dynamics that show up in their data. Our aim is to use advanced analytics to uncover the story in the data that is specific to the organization, a story that leaders can understand and that will compel action. Our disciplines provide the basis for interpreting empirical findings and developing those all-important stories. These interpretations matter because they may have very different decision implications, as our regional bank quickly discovered.
Prometheus: What else makes a good workforce analytics group?
Haig: A good workforce analytics group separates reporting from intelligence. I feel very strongly about this point: Do not have these activities done by the same people or in the same unit. Reporting is about capturing and disseminating basic facts, for instance, voluntary turnover rates, spans of control, employee demographics, etc. It’s very transactional in nature and tends to be demand-driven. Once you do a good job at that, there’s ever more demand for it.
So, colleagues in the business call your team regularly for this data point or that data point. And often they need it by the end of the day. Everything else stops as the data request is fulfilled. This is a very different activity from undertaking deep reflective analyses that are guided, not by a request for data, but by the need to understand what matters, to uncover the risks and the opportunities emanating from the human capital side of the business.
Such proactive inquiry goes beyond showing what is happening in the workforce to explain why they are happening. It is a more complex exercise requiring very different skills, backgrounds and motivation. Reporting is important but explaining and predicting is what most differentiates effective analytics departments from the run-of-the mill functions.
Sadly, the analytical gets crowded out very quickly by the transactional. The good news is technology can really help here because it is very simple to create all the reporting you need using technology. Reporting is amenable automated processes. I recommend setting up reporting that way and periodically revisiting the content based on what the more advanced workforce intelligence modeling reveals.
Prometheus: What about AI technology?
Haig: Clients often ask about AI. They want to know to what extent machine learning can be deployed. I am dubious about this, at least at this point in time. I don’t think AI can capture the complexities of interactions within systems of HR practice and workforce events that deliver the most profound insights.
I don’t think AI can as yet embody the kind of interactive testing against stylized models of HC management of the kind I referenced earlier. Maybe one day it will, and our disciplines will be sidelined by mechanized iterative modeling. But, as the character Juba says at the end of the film, Gladiator (and with the same joyful conviction), “Not Yet!”
Don’t read that comment as suggesting I am completely negative on AI. AI can be usefully deployed, for example, in supporting high volume, routinized functions like recruiting where, for instance, it is useful to triage potential candidates from many sources and use algorithms to flag candidates more likely to be quality hires.
Transactional efficiency is very important in such circumstances and data science does bring value to the table. Further, it contributes new methods and algorithms for analyzing data that are different from traditional statistical analyses. If also offers creative ways of displaying data and analytical results.
Prometheus: How has the working virtually impacted the Institute?
Haig: Our clients adapted very quickly to the reality that we wouldn’t be meeting in person for a while, that everything would be virtual. We have seen a big uptick in the degree of exchange with clients. We used to save things for the in-person meetings, those high-intensity sessions where we’d work through findings and implications together. Then the teams would go their own ways until the next round. Now we are having continuous virtual meetings and there is a lot of regular dialogue. In that sense, the network with customers is actually growing.
Clients are interacting with us more and asking for more interim reports. While the efficiency of delivery of a narrow task is much easier, the sheer volume of tasks is greater. Time for the more complex thinking is often reduced. The work to make sense of data and really turn it into intelligence is not as well facilitated by technology as it should be.
There are challenges with virtual work we need to worry about. There is always the question about the diligence of people working remotely? Early on after the onset of COVID-19, I did hear some clients express concern about the potential falloff of productivity with remote working. But this quickly faded. I think many leaders were pleasantly surprised by how productive their workforces remained. You know, most leaders don’t want to get Orwellian and start tracking time on computers or use other such intrusive tactics. Not in a professional environment.
I can tell you not one of my clients has complained of productivity falloff so far. There are a few clients who have raised concerns about maintaining engagement. We may be “meeting” more frequently because of the technologies of remote working but are the conditions that promote high engagement in place or are they diminished and leading to loss of engagement? That’s an open question.
Prometheus: Tell us more about where technology helps your work?
Haig: I think so far, technology is best at supporting the fulfillment of tasks. It has not yet found the way, at least in my experience or what I have gleaned from our client experience, to facilitate the broader exercise of collaborative thinking and problem solving.
There are collaboration tools created to facilitate active collaboration I’ve seen and worked with some of those tools. I’m not convinced they can replace the value of sitting in a meeting room together or walking down the hall to engage with a colleague with whom you’re working to noodle over an issue or idea. In our group, we are very conscious of the importance of team and collective thought. During this period of remote working, we’ve tried to schedule times where people just come together and talk about what they’re doing. Honestly, it is just not the same as when we’re together in the office.
Is the facility of using Zoom, Microsoft Teams, or other such tools putting people on automatic pilot? Is it easier just to be task-focused where you start to look at your job, just as a succession of tasks — deliver them and move on — versus a work process of both delivering and learning from work? Does the learning piece not just come from doing the task but from interactions with others who bring new ideas, new concerns, new perspectives, and capabilities to address the problem you are focused on in ways you never would have thought of before?
These are important questions that will need to be addressed, particularly in organizations that determine, post COVID-19, to continue with remote working to a significant degree. I have seen no compelling science on these questions as of yet. Speaking personally, I find the technology of remote working is conducive to a task view of work. Let’s get the work done. Let’s keep on time. Let’s check in with each other to make sure we’re doing things right and on time. It’s much more programmed and it’s more about activity — sheer volume. Activity can be the enemy of thinking.
In my line of work, thinking should have primacy over activity. Hence, I do have concerns about the shape of work to come if technology has full sway.
We are a non-profit group of senior technologists and business practitioners seeking to improve society, its institutions, and businesses through information technology. We make observations, challenge assumptions, offer insight into technology endeavors by sorting out their promises and pitfalls along the path to successful implementation.
