Integrating utility analysis and workforce strategy research: suggestions for future work

Abstract

Firms ask their employees to perform a wide variety of tasks, often with daunting time constraints. Research on the firm-level impact of these behaviors – including work in utility analysis (UA) and star employees - has a long and fruitful history, rich with managerial implications. In this paper I comment on research by Joo et al. (2022), who advanced the literature on UA and star employees by highlighting the effects of variance in workforce performance on important firm-level outcomes among 824,924 employees, pooled across 206 samples. In my view this literature can be substantially enhanced by incorporating important moderating and mediating variables that have been identified in the HR strategy literature. In addition, this work can also be improved by developing a better understanding of the causal processes through which star employees’ performance helps to execute strategy, and through the development of better workforce analytics. Advancing this line of research will likely require both qualitative and quantitative research as well as extensive case studies on the identification and implementation of effective workforce strategies.

Keywords:

• HR Systems
• Star Performers
• Strategic Human Resource Management
• Strategic Jobs
• Utility Analysis
• Workforce Analytics

A significant and recurring theme in the human resources management (HRM) literature concerns the question of value: Do relatively higher or lower levels of employee performance make a difference for firm-level outcomes, and if so, how big is the effect? And, if the variance in employee performance does matter, then how does the HR function affect the attraction, selection, development, retention, and exit of talent? These issues are fundamental to the field. If better employee performance does not make a difference, the argument goes, then it is unlikely that HRM policies and practices will, either.

These concepts have been of central importance to both scholars and practitioners for over 100 years (Munsterberg, 1913). Although early work in this area focused on impact in terms of statistical significance, one of the primary contributions of the utility analysis (UA) literature has been its focus on estimating effect sizes in practical terms (Brogden, 1949; Schmidt et al., 1979). More recently, the literature has focused specifically on the returns from investing in ‘star’ employees, that is, the highest performers (Aguinis & O’Boyle, 2014; Groysberg et al., 2004; Groysberg & Lee, 2008, 2009). While the literatures on UA and that on star employees have developed somewhat independently, they both focus on the same broad research question: Does variance in workforce performance make a difference, and if so, how big is the impact?

It is in this broader context that Joo et al. (2022) made their important contribution to the literature. Prior work has established across a broad range of contexts, situations, and time periods that better ‘talent’ (broadly defined) is reflected in a wide variety of firm-level outcomes. By combining estimates derived from UA taken from over 200 studies, Joo et al. (2022) increased the precision of the estimates of the firm-level impact of star employees and concluded that ‘the HRM function can add greater financial value as a function of adding more stars’. They also identified nonlinear effects in this relationship, finding diminishing returns at the highest levels of the talent–firm performance relationship.

However, the finding that ‘more talent is better’ raises a number of questions, not only for Joo et al. (2022) but also for the field as a whole. For example, the authors’ finding of decreasing returns in the employee performance–firm performance relationship might mean that the returns to superior employee performance are not as robust as scholars previously found. However, might not the opposite result also be true? Consider: some scholars have argued for the presence of what economists call increasing returns to scale, where the returns to talent are not diminishing but rather increasing (Becker & Huselid, 2006). In this case, it might well be worth the necessary efforts to secure the services of ‘best-in-world employees’ (as many sports fans will argue during free agency or draft season!). Or is it the case that both effects exist in the same firm, depending on factors such as strategy, situation, and work structures? Perhaps most fundamentally, if these effects are so large and easily observed, why do they persist? Why do motivated managers not act to capture these gains if they are so obvious? Are these effects, then, too good to be true, or does widespread investment in top talent represent a persistent and enduring opportunity for firms?

The evolution of my thinking on utility analysis

My own research on the topic of UA began in the mid-1980s. As a graduate student and management consultant at that time, I found the work of Schmidt et al. (1979) particularly appealing. They, along with others, argued that the standard deviation of employee performance (SD_y) averaged between 40% and 70% of employee salary (Schmidt et al., 1979). What does this mean in practical terms? The implications are that increasing employee performance from the 50th to the 84th percentile, for example, would generate between $40,000 and $70,000 (per employee) for an employee with a salary of $100,000. But what do these figures represent? Can we expect these effects to be linear, or do they take some other functional form (e.g. increasing or decreasing returns to scale)? Further, do these numbers reflect gross increases (e.g. sales), or increases net of salary, taxes, and other associated costs? Are they increases for a single year, in perpetuity, or do the returns exhibit some sort of decay or perhaps even an increase over time? If they do persist over time, do these figures represent adjustments for the time value of money (Boudreau, 1983, 1991)? And what about the labor market? Would higher-performing employees not expect to share in much of the value they create, thus potentially reducing the effect size estimates substantially, especially in a tight labor market (Becker, 1989)?

These and many other questions quickly appeared in the literature after Schmidt et al. (1979) published their initial findings. Although both critiques and confirmatory findings proliferated, an important and durable finding persisted: ‘Better’ employees were significantly more valuable than had been previously assumed in both research and practice. This was certainly consistent with my experience that talent was the key defining attribute in organizational success for many firms, but to see it as the focus of such a substantial body of research was a revelation. As a young scholar and practitioner, I found this important and highly motivating for my own program of research.

However, I still had many questions. One of the primary implications of these findings was that high-performing employees—and the employee selection systems that help to bring them into organizations—were substantially more valuable than we had previously understood. But how could these effect sizes be so large and persistent? If the effects were that large, would firms not move to capture these gains, factor-price equalizing them over time? Perhaps not, if there was a substantial market failure in our understanding of the implications of variance in employee performance in organizations, in that managers have misunderstood (and misvalued) the workforce that is all around them. Or might managers know something that scholars do not? Perhaps both could be true, in that managers do not fully understand the potential opportunity associated with hiring better employees, and scholars do not fully understand the practical challenges with capturing these gains. In any case, for me it was an interesting puzzle with important and widespread implications.

In my view, if the information market failure hypothesis was plausible, then educating managers about these findings might be a solution. At that time, I was spending a great deal of time working with managers and leaders on workforce management issues, and I regularly shared the UA findings with them and asked them for feedback. Quite consistently, the reaction would take the form of ‘we believe that higher-performing employees can create a lot more value—but finding and keeping them is exceptionally difficult. There is a big difference between understanding that talent is so valuable and being able to do something about it’.

These experiences fueled my interest in this line of research. At this point it was clear to me that both talent and the firm’s workforce management system matter—in part because there is substantial real-world variance in both. However, as we all know, without variance in a predictor (e.g. workforce or HRM system), there can be no effect in terms of statistical significance: The correlation between any variable and a constant is zero. But just because there is variance in employee performance does not mean that there will be an impact at the level of the firm. So, variance in talent is a necessary—but not sufficient—situation to observe the results I found in the UA literature. At this point, the question for me was: how did of these pieces fit together?

These fundamental questions motivated much of my own program of research starting in the mid-1980s. In fact, one of my first publications in a refereed journal (Becker & Huselid, 1992) focused on using regression to calculate SD_y in a managerial sample as a validity check on the estimates provided in the prior work. What Brian Becker and I found was that the SD_y of employee performance in dollars was 74% to 100% of salary, considerably greater than the estimates provided in prior research. Persuaded that the variance in workforce performance had important implications for firm-level performance, we shifted our level of analysis over the next few years in both independent and dependent variables, focusing on the impact of HRM practices and systems on turnover, productivity, and corporate financial performance. For us, this led to a series of studies focused on the firm-level impact of HRM systems (Becker & Huselid, 1998, 2006; Huselid, 1995).

So, at this point I concluded not only that talent usually matters (though not always), but also that managers do not always appear to behave in ways that capture these gains. Why might managers not understand how talent makes a difference in their own organizations? From my experience, there are many reasons. In contrast to other information management systems that might exist in finance, accounting, or supply chain, there is a paucity of workforce data in most organizations. When these data do exist, they are often linked to HR function activities (in contrast to workforce outcomes), such as cost per hire or benefits as a proportion of revenues, which encourage a short-term and transactional approach to managing the workforce. In addition, accounting standards worldwide encourage expensing such investments as training and development, which, in comparison to investment in plant and equipment, also encourage managers to make suboptimal workforce investments. Expensing workforce investments means that managers are required to pay for these investments in their entirety from the current year’s budget, which increases their costs dramatically compared to investments that they can depreciate over time.

In short, managers often have a clear idea of how much they spend on the workforce but a much less-developed understanding of how those investments create value over the long run. When making decisions about investing in their workforces, managers often must compare a certain cost today with an uncertain return in the future—and because the only outcome data they have is often linked to employee performance appraisals (which themselves often suffer from range restriction and other issues). As a result, HR investments are often a difficult ‘sell’ to managers. Thus, even though talent is the most expensive investment most firms make (typically between 50% and 70% of their revenues on direct and indirect workforce costs), they often do not understand, measure, or manage this investment well.

In practical terms, what can we do about this situation? For my colleagues and me, a first step in the process was for firms to develop a bespoke information system that helps them to collect data on, manage, and improve their workforces, with a particular focus on strategy execution. This led to fieldwork in a wide variety of firms and The HR Scorecard: Linking People, Strategy, and Performance (Becker et al., 2001), a book that focused on designing and implementing measurement systems for the HR functions. Our continued work led us to write The Workforce Scorecard: Managing Human Capital to Execute Strategy (Huselid et al., 2005), which broadened the focus to the workforce and strategy execution with a particular focus on creating information systems for line (as opposed to HR) managers. These experiences led to our focus on workforce differentiation and analytics highlighted in The Differentiated Workforce: Transforming Talent Into Strategic Impact (Becker et al., 2009) and subsequent papers.

In The Differentiated Workforce, we argued that the most successful and profitable firms manage their workforces with a portfolio approach, placing the highest-performing employees in the highest-return strategic positions. We found that less successful firms, in contrast, invest in talent by hierarchical level. That is, the greatest investments in talent (resources during the selection process, training, developmental opportunities, and compensation) tend to occur for the most senior employees. Because these investments are often quite successful, what this means in practical terms is that the variance in workforce performance tends to go down when the employees’ level in the organization is higher—meaning both that the opportunity and relative ease of impact is greater at lower organizational levels. Put plainly, it is often much easier to improve the performance of a high school sports team than that of Olympic-level athletes in the same sport—yet this is exactly what firms do with their workforces: They focus on the top, not the bottom of the performance distribution. Firms therefore tend over time to underinvest in strategic work and overinvest in nonstrategic work.

Given this, I think of my own research program (which continues to this day) as one long series of investigations in which I try to better understand how variance in employee performance ‘matters’ and to focus on the puzzle: If talent is as important as the literature suggests, why do firms not do a better job of managing it? As of this writing, talent shortages affect businesses throughout the world. Especially in this context, my view is that both talent and the organizational management systems that help to generate it are increasingly important and powerful assets. It is also my view that many firms both underinvest and mis-invest in talent. However, capitalizing on these potential returns will require us to think carefully about the types and location of talent that make a difference.

Suggestions for future research

In the spirit of helping to advance this literature, I offer several observations that I hope will inspire future research. I do not intend to criticize Joo et al. (2022) paper; my aim instead is to make helpful suggestions to advance research on star employees and UA more generally. I begin by reiterating the central question: If the returns to star employees are as large as the results suggest, why do these effects persist over time? Why do managers not do a better job of managing their workforces to capture these gains? I propose several fundamental reasons.

First, attracting, selecting, developing, and retaining world-class top talent is much more complicated and difficult than the UA literature implies.

Second, their location within the firm heavily influences the value of high-performing employees. There are some roles—we call them strategic jobs—that offer substantial opportunities, whereas the upside opportunities in many other roles are limited (Becker & Huselid, 2006; Becker et al., 2009). Because most of the UA literature reported findings averaged across all jobs, the average effect sizes cannot necessarily be generalized and extrapolated across the firm.

Third, the upside potential of employee performance reflects increasing returns to scale in some jobs, whereas it is asymptotic (or perhaps even negative) in others. The impact of poor employee performance on organizational success also reflects this asymmetry. Put differently, there are many roles that can destroy wealth in organizations but few that can actually create great wealth. Identifying and managing these roles are a key component of strategic workforce leadership that the UA literature has yet to reflect.

Fourth, the broad implications of these points for Joo et al. (2022) paper is that it is, in my view, essential to drill down and provide more fine-grained detail if we want to further improve the estimates of the financial returns to hiring star employees. Hiring a highly skilled star employee with rare, valuable, and nonsubstitutable skills (e.g. a natural language processing [NLP] programmer) could have significant positive returns in one firm but negative returns in another firm given differences in firm strategy and the substantial resources required to attract, select, develop, and retain such an employee. Moreover, the returns to NLP programmer ‘stardom’ could likewise differ within a given firm, depending on the importance of the project the programmer works on. Although Joo et al. (2022) do account for occupation in their analyses, my point is that job title or occupation is unlikely to contain enough information to parse the variance of importance here, and scholars may have to incorporate such concepts at the point of research design prior to data collection.

Integrate the literature on HR strategy, star performers, and utility analysis

Beginning in the early 1980s, the scholars writing about HR strategy simultaneously attempted to link their work with that on UA (how do workforce needs differ by firm strategy?) and began to develop an understanding of the financial returns for firms that were able to affect this difference (which HR practices make a difference?). Whereas the UA literature was conceptually and empirically based in psychology, the HR strategy literature was generally grounded in HRM and strategy—although there clearly was some overlap. In the main, scholars in this line of research concluded that the financial returns to investments in what have come to be known as high-performance work systems are both economically and statistically significant (Becker & Huselid, 1998, 2006; Combs et al., 2006; Huselid, 1995). Many of the conclusions derived from this line of research are entirely consistent with familiar principles of sound HRM. These include (a) careful selection and hiring consistent with the firm’s competitive strategy and operational goals, (b) reward systems that reflect the elements of successful strategy implementation in appraisal systems and compensation, and (c) development strategies that emphasize training and performance management systems guided by business objectives (Becker & Huselid, 1998; Huselid, 2015).

One way to think about the linkages between these two lines of literature is to say that the UA literature focuses on the returns to variance in employee behavior, whereas the HR strategy literature focuses on the returns to investments not only in specific types of employee behaviors but also HR policies and practices. The two lines of research are linked in that, if there was no return to better-performing employees, there would presumably be no returns to the HR systems that manage those employees.

What are the implications of the HR strategy literature for the UA literature? One of the key contributions of the HR strategy literature is that different strategies will require different behaviors to execute them, and eliciting different behaviors will require different HR policies and practices. The implication is that, if implemented effectively, HRM systems should likewise differ by strategy. Extending this logic to the firm and the employee, we might expect different jobs (and employees) to be differentially important across firms. One might ask the question this way: Do firms really need—and can they truly afford—world-class talent in every position throughout the organization? If yes, then we might want to adopt the UA findings throughout the firm. If not, then we need to reconsider some of the recommendations emanating from the UA literature. It is especially important to focus on HR systems (not practices) and on jobs in addition to employees.

Focus on the HR system, not the HR practice

One defining characteristic of the HR strategy literature has been its focus on HR systems, not individual HR policies and practices. Why is this important? Historically, the UA literature has focused on the returns to employee selection systems, and many of the papers have modeled the financial impact of various levels of selection predictor validities (e.g. what would the impact be of increasing predictor validity from .3 to .6?). However, one fundamental problem relates to incomplete model specification, specifically missing variables in the operationalization of the HR system construct. For example, employee performance is not solely a function of selection validity, but is also strongly influenced by onboarding, training, performance management, compensation, culture, and strategy. Scholars (Huselid, 1995, Becker & Huselid, 1998) found that there are strong system level HR ‘quality’ effects in organizations, and that a focus on, for example, selection predictor validity is likely also to include the effects of the ‘other’ HRM practices. So, ascribing an effect to a selection system is likely to be overstated, because the coefficient on selection will also carry significant information within it about the quality of the rest of the HRM system.

In addition, my colleagues and I also found that there were strong complementarities and systems effects in most organizations, such that a big increase in predictor validity (as modeled in the UA literature) without a concomitant increase in the quality of the ‘rest’ of the system is not likely to generate the expected returns. Put differently, hiring great people and then exposing them to subpar training, development, performance management, and compensation systems are not likely to be an effective strategy. The solution to this problem is to adopt a systems perspective and, whenever possible, provide estimates of HRM system utility, in contrast to utility estimates for individual HRM practices. So, whereas the UA literature in general and Joo et al. (2022) findings in particular focused on the returns to selecting or obtaining star employees, capturing the returns from star performance will, in my view, require a much broader intervention than simply changing the selection protocol. Put differently, although Joo et al.’s findings might represent the financial impact of the entire HRM system, they are likely to overstate the returns to selection.

Explore strategic jobs: the differential returns of employee performance by firm strategy and job

A second key contribution of the HR strategy literature to UA research is the finding that the returns to employee performance can and often do differ dramatically by job and level with the organization. Thus, estimating the ‘average’ SD_y in a given organization will likely include substantially different variance estimates, and this will have important implications for managers hoping to adopt these findings in their own organizations. My colleagues and I would suggest that most firms neither need nor can afford world-class talent throughout the organization. Why might this be the case? Perhaps the primary reasons relate to our definitions of world-class talent as well as the costs in terms of time and money associated with the attraction, selection, onboarding, development, and retention of this caliber of employee. Much as in real estate—where price is largely a function of ‘location, location, location’—some ‘locations’ (i.e. jobs) are much more valuable than others.

For example, high and low levels of program manager performance (a common job in many organizations) can have dramatically different values in different firms, sometimes even in the same industry. Moreover, the same role can have very different values depending on the location within the firm. For example, my colleagues and I performed a detailed study of the mortgage processing operations within a large bank. We found that the returns to investments in loan officers (of which the bank had many hundreds) differed dramatically depending on whether they worked in mortgage refinancing or new loan origination. The former was a highly routinized progress, and most of the variance in performance was operational in nature. In contrast, the process for new loan origination was much more idiosyncratic—it was in essence a sales role. There was both a higher variance in employee performance in mortgage origination than in refinancing and greater swings in financial performance. At the level of the organization, however, the employees in both of these roles were called loan officers and were managed in a similar manner. A less fine-grained approach to understanding how these roles contributed to firm success would have missed the most important source of variance in the system.

Thus, my colleagues and I made the distinction between jobs that are important, that is, they are indispensable, and jobs which are strategic, that is, investments in these jobs and improving the performance of people holding them will lead to real, observable improvements in firm-level outcomes. This distinction is meaningful because the relationship between employee performance and firm-level performance is not linear for many jobs. It is determined by both the nature of the work being performed and the distribution of talent in the firm’s portfolio.

Consider how a workforce strategy might differ in the retail industry, even for two highly successful firms such as Costco and Nordstrom (Becker et al., 2009). Both firms are well known, profitable, and well respected in North America and both have high customer satisfaction scores, including ‘buy again’ and ‘recommend to others’ scores. Although their portfolio of products differs, they are both in consumer retail goods. Yet their workforce strategies—and consequently their strategic jobs—are quite different. Costco customers look for low prices, large quantities, and clean stores. Nordstrom sales staff function as personal shoppers who must be knowledgeable, timely, and courteous. Customer satisfaction is a key driver of revenue for both firms, but the drivers of customer satisfaction are quite different, and the strategic jobs and requisite employee behaviors differ as well. A sales associate position is strategic at Nordstrom but not at Costco. Purchasing, supply chain, logistics, and real estate management are strategic positions at Costco but not at Nordstrom. So, each firm’s management systems will need to differ as well—and in fact they do (Becker et al., 2009). The point I wish to make here is that a ‘one size fits all’ strategy is likely to be suboptimal for both firms, as would benchmarking across firms. I hasten to add that a strategic job is not the most senior or visible role. Rather, a strategic job has the dual characteristics of being both high impact (nested in a firm’s strategic capability) and high opportunity (considerable variance). In the context of Joo et al. (2022) findings, if both Costco and Nordstrom code their employees as ‘sales associates’, then a critical source of variance associated with firm strategy will be unobserved. As I argued earlier in this paper, it will likely be difficult to remedy this situation in post hoc or secondary analyses. Rather, this is an issue of important consideration in research design.

How do we integrate these ideas and advance this process? The strategic jobs construct may offer a path forward (Becker et al., 2009). The question that my colleagues and I were interested in answering in the early 2000s was this: How do we help managers narrow their focus to investing in the most important jobs and employees? We focused on the underlying issue of variance in workforce performance, although not in perhaps the conventional sense. We considered strategic jobs to be those nested in a firm’s strategic capability (e.g. new product development) and which exhibited significant variance in performance. So, for us, strategic jobs were important because (a) better performance in these roles has a demonstrable effect on firm level outcomes, and (b) the variance in performance levels offers an avenue to capture improved performance. Consider a counterexample, such as a job with significant potential impact but little opportunity for upside. For example, imagine a drug development team in a pharmaceutical firm that contains the top ten scientists in the field, or an Olympic team comprising the top athletes in the nation. Conceptually, it is quite clear that talent is probably the driving force in performance in each case. However, because there is little variance in the capabilities of team members (i.e. they are all world-class performers), there is little opportunity for improving selection quality to make a difference (although there are many opportunities for training, etc., to do so). Note however, the asymmetry inherent in this relationship: increases in selection quality probably won’t drive improvement, but decreases in selection quality can be disastrous. Conversely, imagine a job with substantial variance in team member performance (i.e. the gap from low to high performers is large) but little opportunity for impact because higher employee performance makes little difference at the firm level. Here, efforts to improve selection systems will not matter much because higher employee performance is not differentially important.

Contrast these examples with the situation where strategic impact is large (better employee performance would matter) and the variance in current workforce performance is large (e.g. the gap between the top and bottom of the workforce is large). Here, investments in better selection systems would likely yield substantial returns and would likewise be worth the discretionary effort and investment of the management team. Strategic positions matter because they offer the greatest potential upside in performance improvement (Becker et al., 2009; Becker & Huselid, 2006; Huselid et al., 2005). Therefore, the intersection of impact and opportunity makes the difference—strategic capabilities provide the context, and variance in performance provides the opportunity. Both are necessary to maximize the impact of workforce investments.

Strategic jobs can appear at any level in the organization but, in our experience, they tend to be the most prevalent in the middle and lower levels of the organization and rarely represent more than 15% of a firm’s jobs. The assertion that a firm’s most senior positions might not be strategic might seem odd, perhaps counterintuitive. However, my colleagues and I have defined strategic positions as being, in essence, investment opportunities. Our reasons are several. First, there is less relative variance in senior roles. Consider the process involved in becoming a senior leader in an organization (or top athlete, or scientist, or musician). Becoming the ‘best’ requires not only innate talent but also substantial experience, and practice. In an organizational setting, this typically means that low performers will systematically leave the talent pool over time (through performance management, promotion processes, and self-selection), and the variance in the pool will decrease over time (although the mean will certainly increase). Consequently, although there are exceptions, our data show that the relative variance in performance is higher at lower organizational levels. This means that variance provides opportunity, and because there is often substantial variance in lower-level positions in organizations, the opportunity for improvement is greater as well. The implication for the UA literature is that a focus exclusively on average firm-level effect sizes will lead to underestimates of the impact of strategic jobs and overestimates of the impact of nonstrategic jobs.

Explore nonlinear returns in the HR–firm performance relationship

As a corollary to the strategic jobs concept, our data also show that the ‘returns to talent’ (i.e. the returns to better employee performance) are not linear, although our findings are quite different than Joo et al. (2022). They found decreasing returns to talent at the highest levels of the distribution of their pooled sample. Our data and experience would suggest that some jobs—what we call strategic jobs—actually exhibit increasing returns to scale (Becker et al., 2010; Huselid, 2018; Huselid et al., 2005). A sports team recruiting the best position player in the league or an auto manufacturer recruiting a top designer might be relevant examples. Indeed, this relationship is complicated; Groysberg and his colleagues found that the effect of recruiting this level of ‘stars’ can only be captured if a team of ‘star’ coworkers is recruited as well, what Groysberg and Lee (2009) referred to as a ‘lift-out’. The relationship between employee performance and firm performance in other jobs (support jobs, in our terminology) quickly becomes asymptotic, leaving little incentive to improve workforce performance because, beyond a certain point, improved performance simply does not matter. Developing a better understanding of these issues is not just an arcane academic issue, but rather has important organizational implications. For example, if Joo et al. (2022) findings were correct, they would suggest that above a certain point, managers should abandon their search for more and better ‘star’ employees. If, as I have argued here, the strategic jobs approach was correct, it would suggest that managers should identify their most important and highest return jobs and staff them with world-class employees. In contrast, other roles throughout the firm might not benefit from better talent. These two approaches represent different ways to manage a workforce, with significant implications for all stakeholders involved.

Link the UA and ‘stars’ literature with the field of workforce analytics

My colleagues and I take the view that the implications of what we know so far strongly suggest that leaders should manage their workforces strategically, and this process should begin with a clear understanding of how value is created for current and potential high-value customers. The next stage in the process is identifying the strategic work and strategic jobs that disproportionately contribute to creating value for those customers. Once managers have identified them, they can then focus on understanding the inventory of talent that is available to them to staff those roles. Only then can managers develop a unique and differentiated strategy for the firm that can help distinguish it over the long run. Finally, metrics (measuring our most important workforce indicators) and analytics (building psychometric and econometric models to improve these measures) can be developed (HR and workforce scorecards). Managers can then design and implement any needed changes in their HRM system, and the findings from the UA literature can be especially salient.

This process is the essence of the nascent field of workforce analytics (Huselid, 2015), which draws on a variety of social science disciplines to focus on understanding, managing, and improving the processes though which talent contributes to organizational success. This includes a focus on levels of workforce success across a wide variety of attributes (e.g. how many assistant store managers are ready for promotion to store manager?) as well as relationships or analytics (e.g. what are the primary drivers of assistant managers’ readiness?).

Conclusion

Firms pay their managers to differentiate, make decisions, and allocate scarce resources, often with daunting time constraints. Research on the firm-level impact of these decisions has a long and fruitful history, rich with managerial implications for organizations hoping to become more efficient and effective and generate meaningful jobs. This research often focuses on the workforce and on developing an understanding of how the variability in talent makes a difference. Joo et al. (2022) advanced the literature on UA and star employees by highlighting the effects of the impact of variance in workforce performance on important firm-level outcomes.

Sourcing global levels of top talent requires significant time, energy, focus, and money. The UA literature highlights the importance of developing a better understanding of the financial returns to this important and scarce resource. In my view this literature can be substantially enhanced by incorporating important moderating and mediating variables identified in the HR strategy literature and by obtaining a better understanding of how star employees’ performance helps firms to execute strategy through the development of better workforce analytics. Advancing this line of research will likely require both qualitative and quantitative research as well as extensive case studies on the identification and implementation of effective workforce strategies (Becker et al., 2009; Huselid, 2015, 2018). Joo et al. (2022) provided an important contribution to our understanding of the role of employee performance in firm success. However, in my view it will be important to design studies that incorporate the concepts of both HR strategy and UA from the outset, because it is difficult to test these concepts using archival data.

I began this paper with the personal observation that research in UA helped to provide the genesis of my nearly four-decade interest in the role of talent in organizational success. Then as now, I am convinced that most firms underinvest (and mis-invest) in people, influenced not only by the literature but also by some incorrect assumptions about the nature and role of talent in organizational success. In this paper I have attempted to highlight what are, in my view, the most promising future directions and opportunities. Both the UA and HR strategy literature would benefit from a more refined understanding of the processes through which the workforce creates value in organizations, and I have attempted to share some options here.

Disclosure statement

No potential conflict of interest was reported by the authors.