Employee engagement is a concept of considerable interest to organizational stakeholders and HR researchers. Over the past decade, it has been hard for practitioners to ignore mounting research demonstrating the business relevance of work attitudes such as employee engagement (e.g., Macey and Schneider 2008) and satisfaction (e.g., Judge et al. 2001; Whitman et al. 2010). Earlier this decade when the economy was stronger, it was routine—almost trendy—to measure employee attitudes using organizational surveys. Unfortunately, many in the business world have recently noticed a different trend in the current economic climate: employee surveys are often among the first programs to be postponed or removed during budget cuts and cost containment measures. In order to mitigate this knee-jerk (and we argue, often unwise) reaction, it is imperative to continue to extend research so that we can better understand the real-world impact of an engaged and satisfied workforce on an organization’s bottom line. This will involve moving beyond assessments of construct validity and into practical ROI and business impact analyses that resonate with key organizational decision makers.

In this article, we first briefly review the concept of employee engagement and its relationship to employee satisfaction. Second, after establishing the general construct of emphasis, we turn to reviewing existing research—along with drawing upon real-world examples and data across a variety of organizations—on the business value of an engaged workforce. Third, using this past research as a baseline, we highlight why cutting an employee survey program during a business downturn can be shortsighted and actually the worst time to do so. Fourth, we offer practical applications for HR managers, including providing guidance on the types of questions that should be asked in an employee survey and focusing on those that result in the most actionable feedback. We conclude this paper by laying out an agenda for future research aimed at helping bridge the academic-practitioner divide.

Employee Satisfaction and Engagement

Until the last decade or so, researchers have in large part focused on employee satisfaction, whereas the business arena has adapted this to the more interpretable and outcome-driven construct of employee engagement. Employee satisfaction involves how individuals feels about their work environment, pay, teammates, etc. (Judge et al. 2001). In contrast, employee engagement is more behavioral in nature and serves as an indicator of performance contributions (e.g., organizational citizenship behavior) consisting of trait, state, and behavioral elements (Macey and Schneider 2008).

Though employee engagement and satisfaction are typically construed as unique constructs, to most managers it appears that the satisfied employee is likely to be engaged and the engaged employee satisfied. This is also consistent with the findings of Harter et al. (2002), who reported a .91 true score correlation between employee satisfaction and engagement. Regardless of which convention one adheres to, it is important to measure and gauge employees’ viewpoints on the work environment. Although we specify how different types of survey questions can be utilized and acted upon, our main argument is that in most instances a variety of employee attitude surveys can provide insightful data and is almost always better than no survey. For the balance of this paper, we refer to employee engagement, but recognize that most of what we discuss applies to any attitudinal employee survey.

The Business Case for Employee Surveys

Employee surveys have evolved considerably and are about much more than simply determining if workers are happy. Rather, surveys are viewed in the business world as an avenue to better understand how to help retain employees and identify other areas where efficiencies can be achieved. The research linking employee turnover to dollar loss is quite compelling (e.g., Kacmar et al. 2006). Meta-analytic work has shown that engagement has a significant impact on a number of business outcomes, including customer satisfaction, productivity, profit, employee turnover, and accidents (Harter et al. 2002). More recent causal investigations have indicated that engagement may have an indirect impact on profit through customer service (Simon et al. 2009). In sum, the implications are clear: higher employee engagement is linked to lower turnover, greater efficiency, better customer service, and improved business performance.

Even in a strong economic environment, a sound business case is a prerequisite for obtaining funding for any initiative. In a difficult economy or business downturn, however, “employee research” can often be viewed as low-hanging fruit when it comes to identifying budget reduction opportunities. As a case in point, Personnel Decisions International (February 2009) recently released survey results based on input from leaders of more than 500 global organizations. Their findings showed a sharp two-year drop in the importance of Talent Management activities related to “Loss of leaders in key areas or insufficient talent to quickly adapt to change”. During “boom” times preceding the current economic downturn, activities in this area were deemed one of the most critical organizational initiatives; during down times, it fell to the very bottom of the list with only 5% of leaders reporting it as critical. We argue that this is a short-sighted approach that can actually open up and expose organizations to further issues, including the loss of critical talent. To support this claim, we present both primary and archival data to convey the importance of an engaged workforce on business performance.

Engagement During Economic Decline

Contrary to popular belief, employee engagement is not guaranteed to wane during difficult economic times. Although much organizational data and business press writings suggest a significant drop-off in overall levels of engagement since the economic crisis (e.g., Matz-Costa et al. 2009), a recent study by Hewitt Associates across 400 plus organizations tells a more complex story. As shown in Fig. 1, though the proportion of organizations experiencing a decline in engagement has increased during the downturn, on average over 40% of organizations still managed to increase engagement (from October 2008 to September 2009). This suggests that positive organizational efforts can lead to an engagement increase even amidst a turbulent environment. In other words, do not throw in the towel and assume that engagement initiatives during this time are unlikely to lead to improvements. As two examples of antecedents, this research has revealed that strong leadership communication and providing cross-training opportunities (which may be more plentiful during a business downturn) have been key drivers of improving engagement. In fact, difficult times actually provide some of the best opportunities for career development, which is another predictor of employee engagement (Hewlett 2010.)

Fig. 1
figure 1

Percent of organizations by engagement score change segment. Based on Hewitt research from over 400 organizations that completed a repeat assessment during the period covering July 2008–June 2009

Full size image

Another issue to consider with engagement is related to how it is measured. For instance, Church and Waclawski (2010) have suggested that there are as many measures of engagement as there are organizations. Measures of engagement also differ in the number of underlying facets of dimensions they contain. The particular model in the figure (i.e., The Hewitt Engagement Model) measures three key dimensions as part of the engagement paradigm: (1) Say, (2) Stay, and (3) Strive. Respectively, these areas measure the extent to which employees speak positively about the organization to coworkers, potential employees, and customers; whether they have an intense desire to be part of the organization; and if they exert extra effort and engage in behaviors that contribute to business success. When we take a more fine-grained look at the dimensions of this engagement measure, some interesting patterns have begun to emerge (July–September 2009). Under normal market conditions, quarterly variations are approximately ±1%. More recently the data has shown that while the “Say” component has stayed within normal bounds, the “Stay” component has increased significantly and the “Strive” component has decreased. Organizations measuring engagement during this economic downturn are finding that their employees are staying put; they are not leaving the organization, nor are they putting in any extra effort. The Corporate Executive Board (2010) reports that they are instead biding their time until the economy improves and then they will turn to actively look for better opportunities. Interestingly, CEB further reports that withdrawal intentions are approximately 15% higher amongst high-potential workers. Although this is just one specific measure of engagement, we believe this data can be used by managers to recognize similar trends within their own workforces and understand what drivers are behind it. An important implication here is that one of the compelling reasons for doing engagement research in times of economic distress is not just to protect your current talent, but to understand how to keep your own top talent and as well as be prepared to attract the top talent from other organizations that are ready to jump ship when the economy improves.

Even if an economic downturn necessitates cuts to a survey program, it no longer needs to be treated as an all or nothing proposition. Many organizations have turned to a variety of sampling approaches in order to continue seeking employee feedback. In the past, organizations have frequently utilized pulse surveys and sampling waves as effective methods to reduce survey costs, while still being able to capture important information on the workforce. However, now that the preferred method of data collection is overwhelmingly through online web sites, simply reducing the sample size will often have little impact on cost since there is very little money to be saved by sending out fewer electronic invitations to take a survey (though, it should be noted, substantial labor savings can be found through less employee time spent completing a survey). The primary way to save money is simply to re-administer the survey without making any changes to the last survey that had been programmed. In effect, one can “push the button again” and thereby save all of the programming and language translation costs. The advantage of getting a full read on the organization remains, albeit, there is an inability to make any needed edits or changes from the last survey format. Eliminating any paper surveys and going with just electronic administration can prevent significant costs due to the added costs of mailing and data entry with paper surveys.

Advice for Practitioners

Many surveys are very effective at measuring aspects of the work environment that can be compared to internal and external benchmarks. Yet still many surveys fall short in identifying actionable insights and solutions for improvement. One way to accomplish this is to arrange survey items around a framework that can be understood by managers. Examples of factors under the framework would include leadership, pay, work balance, and empowerment. Items within each factor need to be directly actionable. Questions such as “I am satisfied with my job” are only providing minimal information. Instead, questions like “I am provided with the technology resources I need to do my job” provide management with insights that can lead to direct action to maintain or improve a specific area. This is one reason we believe it is important to not be fully reliant on “off the shelf” items in an organizational survey: the front-line managers are the ones who know what information is most important to garner from employees. Internally designed survey questions can also be essential for top-management, who can link specific items to strategic plans and organizational initiatives. Furthermore, internal development of items and survey dimensions can lead to important buy-in from senior stakeholders when developed collaboratively.

We also argue that—logistics permitting—employee surveys should begin to make the move from anonymity to confidentiality. This is important so that organizations can better understand how engagement varies based on employee performance levels and other individual-level variables. Historically, this has been difficult due to the strong pressure for organizations to administer fully anonymous surveys. Now that it has largely become common practice to collect data over the Internet, it has become easier to move more towards confidential surveys. Many are already starting to do this by using a “tie-back” approach that links an individual to various pieces of information (e.g., email, tenure, department, supervisor, etc.). All leaders agree that top talent is a premium commodity. To this end, employee data that can be tied to individuals can be extremely informative for purposes of better understanding the mindset of key organizational contributors. At the same time, we must acknowledge that there are some caveats involved with transitioning to confidential surveying (e.g., Edwards et al. 1997). From an ethical standpoint, managers must be held accountable for informing participants of the differences between anonymity and confidentiality. Furthermore, survey instructions should indicate how data will be used, and employees need to be fully aware of exactly who in the organization will have access to their data—and great care needs to be taken to minimize as much as possible both the amount of identifiable information that is collected as well as the number of managers who can retrieve this classified information.

One of the biggest benefits of using a tie-back approach is the ability to view data in novel and important new ways. For example, identifying “High-Potential” employees is very difficult when an organization collects demographic information directly from a worker. Because of inflated perceptions of self-performance, this is not the type of question that can be reliably answered by respondents. However, by using a tie-back approach, high-potential individuals can be identified (typically via performance data stored in an HRIS or similar system) and analyses can be conducted to evaluate and determine what is driving their engagement. It is not at all unusual to find that high potentials value different aspects of the workplace than the rest of the workforce. This is especially true when looking at compensation drivers. Many organizations with very high engagement scores and strong business outcomes subscribe to the approach of providing a disproportionate share of the developmental resources to the high-potential group as they are often viewed as future leaders.

An Agenda for Future Research

Research on employee attitudes and engagement has experienced a renaissance of late. This has been driven by more robust meta-analyses (Judge et al. 2001; Harrison et al. 2006; Harter et al. 2002; Whitman et al. 2010) that have countered earlier work suggesting a limited impact of a satisfied or engaged worker on performance. Thus, research on engagement is in a stage of relative infancy and much more needs to be done in order to understand its antecedents, process mechanisms, and outcomes. Because we view engagement as a multilevel construct, we pose a series of pressing research questions that focus on three critical levels-of-analysis: the individual, the unit, and the organization.

Concerning individual-level research, we believe much more needs to be done to understand and identify the “High-Potential” employees in organizations. It is surprising little research has dealt with the delineation between this group and “average” workers considering retaining high-potential employees is a primary concern of HR managers (Corporate Executive Board 2010). Although we do not wish to redevelop a theory of job performance here, we do want to make a statement based on our anecdotal experiences: job satisfaction may matter more for top workers and engagement may matter more for average workers.

Future research should also focus on asking questions from an individual differences perspective. For instance, we need to better understand the antecedents of employee engagement. Brief’s (1998) model of job satisfaction suggested that antecedents come from both within the employee (e.g., personality) and external factors in the workplace (e.g., quality of leadership). Might we also gain a better understanding of engagement by adopting aspects of this model? Further, what are the person-characteristic factors (e.g., conscientiousness) that predict engagement? To the extent that engagement is attributable to certain personality traits, employee selection processes can be implemented that target these areas. Finally, it will be important to determine if majority–minority differences exist in engagement among legally protected ethnic and gender groups. Engagement is known to be linked to better performance and, thus, to rewards. As a result, engagement among protected groups could be monitored to ensure equivalence and mitigate differential downstream rewards.

At the unit level of analysis, different questions will need to be asked and answered, and unique theory will need to be developed. Although two recent meta-analyses (Harter et al. 2002; Whitman et al. 2010) have shown that engaged business units are more productive, less is known about how and why engagement has such an impact. Attitude-behavior theory (e.g., Ajzen 2005) and human relations theory (e.g., Likert 1967) hold some promise in this area. For instance, Whitman et al. (2010) found that unit-level satisfaction had an impact on unit-level citizenship behaviors, which in turn had an impact on performance. Perhaps separating unit-level behaviors (e.g., efficiency, customer service, turnover) from more distal-oriented criteria like results (e.g., profit) may provide fruitful avenues for research. Scholars could then examine process models that investigate relationships such as: unit-level attitudes → unit-level behaviors → unit-level results. A focus on the role that leadership has in influencing collective engagement appears to be another area of promising research.

We also need a better conceptual understanding of the unit-level engagement construct. For example, most researchers and practitioners conceptualize unit-level engagement using the unit’s average score. However, collective constructs may be represented using a variety of strategies, including dispersion (e.g., standard deviation), highest/lowest score in the group, and median. Examining the extent to which engagement is shared among the work group (e.g., variance, skewness, and kurtosis) and how this “sharedness” influences business criteria are also pressing issues.

Finally, we would like to address some engagement issues from a macro- and organizational perspective. Toward this end, evaluating how engagement varies at organized labor locations compared to non-organized locations could become increasingly important as labor movements such as the proposed Employee Free Choice Act (EFCA) legislation continue to be pushed. We believe engagement surveys can be viewed as a proactive mechanism to identify areas of focus to help predict the likelihood of labor organization activity. Survey providers have already reported an increase in surveys designed to tap into this area and measure union vulnerability.

The issue of time—both contextually and linearly—also needs to be addressed. As far as context goes, job satisfaction may be less important than engagement during a crisis because many employees are simply happy to have a job when unemployment is above 10%. As shown in Fig. 2, this is seen very clearly when looking at differences in engagement across geographies. For example, Latin America often has much higher average engagement scores than the rest of the world. Qualitative research has revealed that in many of these Latin American countries where unemployment is always in double digits, to be employed is to be engaged. If an individual loses his or her job, there is not a very strong likelihood that he or she will be able to regain employment.

Fig. 2
figure 2

Employee engagement by region (based on average score across organizations)

Full size image

Finally, in the business world, we often hear the phrase time is money. Thus, it is also pertinent for researchers to ask questions about how frequently surveys should be administered (e.g., quarterly, yearly) as well as the linear question of how long it takes for an employee’s (or work group’s) improved attitude to be reflected in increased performance or financials. Again, we feel that these issues may highlight the nuanced differences between engagement and satisfaction.

Conclusion

In this article we sought to emphasize that measuring workforce attitudes is a business imperative. Mounting research is making it clear that an engaged workforce is a more productive workforce. Therefore, it becomes a necessity to understand what drives and leads to employee engagement. Above, we highlighted why this holds true during economic downturns, provided specific actions that organizations can do in order to continue to measure engagement, and provided suggestions towards decreasing the cost of surveys. Finally, we believe that significant research opportunities exist with employee engagement and we provide researchers and practitioners with guidance for extending research in this important area.