Strategic Data Collection, Analysis, and Organizational Context
Manager Green is frustrated. One of his primary performance goals for the year is the effective retention of key talent. Following Manager Savvy’s suggestions, Green collected extensive data using a broad decision frame to understand turnover drivers. He also developed a new formula for costing the impact of each instance of turnover, and has data demonstrating that he could save the organization millions of dollars per year by implementing a few straightforward initiatives that would cost at most a couple of hundred thousand dollars to launch. The estimated return on investment was almost immediately positive, and grew at a rapid rate over the next three to five years. When he presented his data and plan to Manager Savvy, she was impressed! However, the presentation to the executive team to get budgetary approval had not gone quite as smoothly. The CFO, in particular, questioned the cost formula developed by Manager Green and the likelihood that the interventions would work as planned. Based on the CFO’s own estimate of turnover costs, plus the opportunity to hire replacements for departed workers at lower cost because of high unemployment rates, he requires significantly more evidence before approving the expenditures. Apparently, nothing is as simple as it first appears.
Kernel of Truth
Sometimes retention is simple. For example, we know that attitudes such as job satisfaction and organizational commitment tend to be related to turnover decisions. Other things being equal, work units with higher satisfaction and commitment will tend to have lower turnover rates, and individuals with higher satisfaction and commitment will be less likely to quit. Thus, paying some attention to job attitudes is almost always a good idea. However, dissatisfaction and low commitment do not always lead to turnover. Even when they do, this knowledge is not as useful without knowing what organizational or managerial factors are leading to negative attitudes in a specific context. Further, without a frame of reference for thinking about the impact of turnover, it is difficult to determine the appropriate resource level to invest in changing attitudes or improving retention. Developing an effective evidence-based retention strategy requires considerably more fine-grained data collection, analysis, and interpretation.
What the Research Says
Research suggests that a strategic evidence-based approach to retention management requires three broad foundations: (1) developing a shared understanding of turnover among stakeholders; (2) developing knowledge of underlying turnover principles and cause–effect relationships; and (3) diagnosing and adapting to a particular organizational context.1
Developing a Shared Understanding
Developing a shared understanding is critical because failing to do so makes it very difficult to design and implement an effective retention strategy. For example, if an HR manager believes current turnover rates are problematic, but the CFO believes the turnover rate is appropriate and an opportunity to reduce labor costs by hiring less expensive replacements, then it is unlikely organizational leadership will be able to agree on the appropriate investment to make in retention. Thus, care must be taken to accomplish three goals: carefully define turnover; develop a consensus for costing turnover; and develop a consensus concerning retention priorities. As described in Chapter 1, there are many types of turnover, and they don’t all have the same retention implications. Organizations need to go beyond simply measuring turnover rates and beyond simply classifying instances of turnover as voluntary or involuntary. Instead, managers need to define how voluntary, avoidable, dysfunctional, and pivotal each instance of turnover is for the organization. To be able to do so, organizational leaders need to develop agreed upon definitions of these terms in a particular organizational context and provide managers the necessary tools (e.g., training, data systems, and resources for third-party exit interviews). Having these data available enables organizational stakeholders to form a consensus as to the extent turnover is an important issue to address, the appropriate investments to make in retention strategies, and how to focus those investments.
Similarly, organizations need to develop a consensus for costing turnover. As described in Chapter 1, there are numerous direct and indirect, tangible and intangible costs associated with turnover. Savvy managers need to know how much each instance of turnover costs the organization, requiring the development of a formula that accounts for the varied costs that may be associated with turnover. The key element of this costing formula needs to be that stakeholders throughout the organization agree that it creates a reasonable estimate. However, even in organizations that track turnover costs, the most common approach is to develop a single cost formula and apply it across all situations; thus, the average cost of turnover is multiplied by instances of turnover over a given time period to produce a turnover cost estimate. Unfortunately, this simple approach ignores the likelihood that turnover costs vary by position or skills, and that some jobs are likely more pivotal than others. Turnover costs will vary depending on who is turning over—losing higher performing, higher quality, more pivotal, or more difficult to replace employees costs much more than the inverse, and one size fits all costing formulae don’t capture this. Organizations are likely to develop relative priorities concerning the costs associated with turnover in different parts of the organization. Thus, the turnover costing formula needs to be flexible enough to recognize and account for these variations.
Developing Knowledge of Underlying Turnover Principles and Cause–Effect Relationships
Another key to strategic evidence-based retention management is to help managers develop systematic knowledge about how and why individuals make turnover decisions. This is critically important because, as we have discussed throughout this book up to this point, managers left to their own devices often develop misconceptions about the processes and drivers underlying turnover. These misconceptions do not take root because managers are apathetic or unintelligent. Most managers are smart individuals who care a great deal about their organizations and the people in them. However, facing continuous organizational pressures and information overload, managers are subject to decision biases such as the sampling bias described in the Introduction. For example, one or two highly salient examples from a manager’s career of losing a valued employee who leaves to take a higher paying job elsewhere can lead many managers to conclude that pay dissatisfaction is a primary driver in most turnover cases. However, we showed in Chapter 2 that this is a widely held managerial misconception.
How can managers develop systematic evidence-based knowledge about turnover and use this knowledge to aid decision-making? We suggest two primary strategies. One is for managers to learn all they can about the research evidence concerning turnover by reading systematic reviews. It is not necessary to read all of the research studies individually: there are numerous helpful systematic reviews available. One resource that summarizes much of this information is, of course, this book. There are also useful systematic reviews that can be found in key journals that bridge research and practice (e.g., Academy of Management Perspectives, formerly Academy of Management Executive) and publications of professional associations (e.g., SHRM Foundation).2 These resources can provide managers a common frame of reference as the research evidence concerning processes and drivers underlying turnover decisions.
The other is for managers to make themselves aware of key biases that prevent them from taking a broad decision frame when faced with understanding turnover. As described earlier, managers are subject to a number of biases when attempting to draw conclusions about important events, and also tend to focus on easily accessible symptoms and solutions to minimize uncertainty.3 Having a broad understanding of the evidence regarding turnover processes and causes can help managers to consider multiple issues and alternatives, not only the first or most salient to arise.
Diagnosing and Adapting to a Particular Organizational Context
Although understanding the underlying principles and cause–effect relationships that drive turnover decisions in general is an important foundation, effective management of turnover requires an additional type of evidence: the nature of turnover in a particular organizational context. That is, beyond understanding, for example, the roles of job dissatisfaction and shocks in turnover decisions, managers need to diagnose what exactly may be driving dissatisfaction in their organization and what shocks are most prevalent and important, and then adapt their knowledge of underlying principles to their particular situation. For example, one organization might find high turnover among hourly employees in a slow-growth market driven largely by a mismatch between expectations and reality; another might find high turnover among experienced professionals in a rapidly expanding market driven largely by disillusionment with career paths. The retention implications of these two examples are obviously quite different. The steps to context-specific evidence-based retention strategies include: thorough ongoing turnover analysis; interpreting turnover analysis through the lens of organizational context; and using multiple data collection strategies to diagnose the primary turnover drivers in a particular setting.
Turnover Analysis: Rates, Costs, and Functionality
Conducting a thorough turnover analysis helps to diagnose the extent to which turnover is a problem, provide data as to where the most important turnover issues reside, and encourage a broad consideration of turnover costs, rates, and functionality, not merely the most recent or most salient concerns (e.g., exit interview data about pay dissatisfaction). Overall turnover rates should be measured. However, it is more useful to also measure these data in terms of key categorizations that can help managers pinpoint the exact nature of turnover in their organization. Some key categorizations can include types of turnover (e.g., functional–dysfunctional; avoidable–unavoidable; voluntary–involuntary), job level, job type, performance level, geographic location, work unit, replacement costs, or demographic characteristics. Breaking the data out in this fashion enables managers to evaluate the extent to which turnover rates in different locations or among particular types of employees may be especially problematic. Organizations able to provide managers with real-time dashboards regarding these data will enable quicker adjustments and responses to potential trouble spots.
The next element is to integrate data on turnover rates with consideration of turnover costs and functionality. As described earlier, it is important to develop cost formulae that enable the calculation of total turnover costs and costs per incidence of turnover. Retention initiatives typically require investment of time, money, or other resources. Agreed-upon metrics for assessing turnover costs enable managers to strategically design and target interventions based on the best available evidence. At this stage of analysis, it is also particularly important to incorporate how dysfunctional departures are for the organization. Not every employee is of equal value to the organization, and not every position is equally pivotal to organizational success or equally difficult to replace. Recall our discussion of pivotal individuals and positions. In less pivotal positions, improving retention beyond a certain threshold may result in diminishing marginal returns, whereas, in more pivotal positions, improved retention continues to have a significant impact on organizational outcomes. Thus, a thorough ongoing turnover analysis that presents managers with credible evidence about turnover rates, costs, and functionality provides a critical foundation for strategically managing turnover.
Interpret Turnover Analysis Through the Lens of Organizational Context
It is not sufficient, however, to consider turnover data in isolation. Turnover rates, in particular, can be very misleading without a careful consideration of context. We have worked with organizations that would be thrilled to reduce their annual turnover rates in some positions below 100%, and with others gravely concerned because their annual managerial turnover rate was approaching 5%. The same turnover rate that is a source of concern in competitive labor markets or during periods of rapid expansion, might be considerably less of a concern during times of retrenchment, contraction, or high unemployment. The impact of turnover rates can vary depending on the industry, labor market conditions, quality of the current workforce, growth of the company and strategic plans for the organization. Therefore, the next step is to interpret turnover analysis data through the lens of your particular organizational context, considering past, present, and future conditions and trends both internal and external to the organization.
Benchmarking and needs assessment are two methods for assessing turnover data relative to internal and external circumstances. Benchmarking entails evaluating turnover data compared to both external competitors and internal historical data. For example, external benchmarking enables managers to determine whether a given turnover rate puts them at a competitive disadvantage with competitors in terms of talent and labor costs. An organization that has a significantly higher turnover rate than competitors, particularly among highly valued or difficult to replace employees, may be experiencing higher turnover-related costs, lower productivity, or service decrements. Conversely, evidence that turnover rates are consistent with or even below competitive norms could suggest that retention resources would be more strategically placed elsewhere. Internal benchmarking entails comparing turnover data to organizational trends over time. Often, increasing turnover signals greater cause for concern than stable or decreasing turnover. This may especially be the case if turnover rates are increasing sharply, or if turnover is increasing among particularly valued groups of employees or in particular locations. However, there may also be cases where stable or even decreasing turnover rates could be problematic, such as when the organization is attempting to downsize through natural attrition. Therefore, benchmarking and needs assessment data need to be integrated with consideration of organizational conditions and plans.
Needs assessment involves the consideration of future labor demand and availability based on internal and external trends. Internal needs assessment considers current and future organizational operational tactics and strategic plans that may influence the supply and demand for human capital of particular types. Some organizational strategies would be likely to increase demand for human capital in general, such as rapid growth in particular locations, growth in particular market locations, or for particular types of human capital. On the other hand, some organizational strategies would be likely to decrease demand, such as downsizing, retrenchment, outsourcing, or offshoring. The same turnover rate would tend to be more problematic and warrant more attention under expectations of increasing demand compared to decreasing demand.
External needs assessment considers labor market conditions, economic conditions and trends, and industry trends as they affect supply and demand for human capital. Some trends are likely to increase demand or restrict supply of human capital desired by the organization, such as industry growth, aging populations, shrinking labor forces, or migration of workers out of certain fields or educational programs. Other trends would tend to decrease demand or increase supply of human capital desired by the organization, such as industry contraction, growing labor forces, growing immigration by individuals with relevant skill sets, or migration of workers into relevant fields or educational programs. Trends that increase demand or restrict supply would tend to make a particular turnover rate potentially more problematic and in need of additional attention. Global companies may even desire functional turnover in certain low growth markets while they seek to reduce turnover in other high growth markets.
Taken together, extensive turnover analysis interpreted in light of organizational context provides extensive evidence concerning the extent to which turnover is a problem, and which incidences of turnover in terms of job types, locations, positions, or individuals are most in need of intervention. Further, the foundation of evidence on turnover processes and cause–effect relationships established earlier provides broad evidence-based principles for managing and reducing turnover in general. However, the missing piece remains context-specific evidence on why valued employees are leaving your organization.
Collect Data to Diagnose Cause–Effect in a Particular Context
There are myriad retention strategies managers could pursue, many of which are discussed in detail in subsequent chapters. However, managers generally have finite resources to devote to retention and must make strategic decisions about which initiatives and strategies to pursue. Further, some strategies are broad-based systemic strategies directed at the entire organization, large subsystems, or large subpopulations; examples may include improving the work environment, changing reward structures, or incorporating retention-related criteria into recruitment and selection processes. Other strategies are targeted at specific job types, work units, locations, or even individuals; examples may include developing a new career path for a particular job, or designing tenure-based incentives for particular individuals. Managers need context-specific evidence of the drivers of turnover in their organization. These data can come from a variety of sources including exit interviews, post-exit surveys, current employee focus groups, survey linkage research, predictive survey studies, and qualitative research. In most cases, multiple data collection strategies are required.
Exit interviews are widely used to collect data on why employees are leaving the organization. Often, a supervisor or HR manager interviews or surveys departing employees and asks questions about their primary reasons for leaving. Some organizations use third parties and collect these data after some time has passed in an effort to get answers that may be more honest or objective. Exit interviews serve at least three important functions: they provide immediate data on why employees are leaving; they can be used as a reactive retention strategy to attempt to convince departing employees not to leave; and they serve a relationship-management function in ending the employment relationship on a positive note.4 However, exit interview data is subject to at least two major limitations. One is that many departing employees are more likely to identify external factors such as a new job and less likely to focus on internal factors such as poor management.5 For this reason, some organizations rely on post-exit interviews or surveys, often conducted by neutral third parties. This post-exit data may be more open and honest than traditional exit interview data.
The other major limitation is that relying solely on exit interviews with departing employees is subject to bias from sampling on the dependent variable as described earlier. That is, it is difficult and often misleading to focus solely on data from leavers without comparing it to data on employees who do not leave. Thus, it is important that managers not rely exclusively on exit or post-exit interview data in developing retention strategies. However, given the multiple functions associated with exit interviews and their widespread use, we can offer some evidence-based guidelines for improving exit interview data: use neutral interviewers; train interviewers; develop a structured exit interview; emphasize confidentiality to the extent possible; and integrate exit interview data with evidence and data from other sources.6
To overcome the limitation of collecting data only from leavers, another data collection strategy is to conduct interviews or focus groups with current employees. These methods enable the organization to collect data from a representative sample of employees focused on identifying reasons employees have considered leaving, reasons they have stayed, and factors they consider most important for the future. These methods overcome the limitation of sampling on the dependent variable, provide new perspectives on why employees stay, and may enable managers to identify potential trouble spots earlier. These methods can also be focused on those employees the organization is most interested in retaining because of job type, location, skills, or performance. Similar to exit interviews, these interviews or focus groups should also be conducted by trained neutral parties and utilize a structured format.
Although useful, interviews and focus groups are time-intensive and typically include only a small subset of employees. Many organizations collect survey data in order to incorporate the perspective of many, if not most, employees. Survey linkage research can be used to link such survey data to retention strategies. The steps in survey linkage research typically involve identifying employee attitudes and opinions of interest; designing a questionnaire; measuring employee attitudes and opinions via anonymous survey; aggregating these responses to the business-unit level; and assessing the relationships among these employee responses, turnover rates, and other important outcomes such as revenues, sales, profits, and customer retention. Linkage research of this nature provides three key types of evidence. One, it provides data about the drivers of turnover at the business-unit level. For example, even if the employee survey uncovers that many employees are dissatisfied with their pay and their promotion opportunities, linkage research may show that only promotion opportunities are correlated with turnover rates. Two, it provides data that can be examined in terms of multiple subgroups that may be of interest, such as by location, manager, job or position, demographics, performance levels, or other factors. Such data may enable managers to develop more targeted retention strategies. Three, it provides data linking turnover rates to other business unit outcomes such as customer responses or profitability. Such data provides managers with additional evidence as to the extent to which turnover is problematic for the organization.
Although survey linkage research can provide valuable data about associations among many possible turnover drivers and turnover rates at the unit and organizational level, this approach only shows correlation, not causality, and does not help managers predict individual turnover decisions. Another data collection approach is to use predictive individual-level survey data in which individual employee survey data is statistically linked with individual stay or leave decisions over time. A valuable strength of predictive survey research is the specific data provided concerning the strength of relationships between specific predictors and actual turnover decisions in your organization. These data can help managers focus retention efforts and strategies, and may even in some cases enable early intervention to head off some instances of turnover before they occur. One important obstacle to consider with predictive survey research is that it requires a method for identifying individual employee responses so that these responses can be linked to subsequent turnover decisions. Unless there is a strong culture of trust and openness in which employees feel safe providing open and honest responses, it may be necessary to use outside researchers or consultants to ensure that employees can respond candidly without fearing reprisals from their manager or from the organization. In some cases, organizations use third party survey service providers that contractually preclude giving individually identified responses to the organization. In such cases, it may be possible to have an external researcher gather the data from the survey provider and the organization to analyze the individual level linkages and generate findings for the organization while still maintaining employee anonymity.
An alternative data collection strategy incorporates in-depth qualitative research that attempts to uncover the rich context and complex processes involved in making turnover-related decisions. Qualitative research involves collecting extensive narrative data from employees, often over multiple time periods, and looking for key patterns or themes that emerge in the way particular employees in a particular organization think about their jobs and organizations. Although often resource intensive, such methods can sometimes reveal issues or relationships not captured by survey research. Examples may include conducting and transcribing the content of repeated interviews with representatives of key employee groups; interpreting the content of employee journals or diaries in which employees are asked to describe their thoughts and feelings about the employment relationship; or experience sampling methods in which employees are asked to respond at regular intervals to short prompts asking about their experiences, attitudes, and thoughts about leaving at particular points in time. This approach provides great depth of information but conclusions are limited to the perspective of the people involved in the observations and interviews. For this reason qualitative data collection is often used in conjunction with quantitative survey research to confirm that identified themes hold up across the organization.
Evidence-Based Management Implications
In summary, strategic evidence-based retention management requires managers and organizations to collect extensive data about the nature and drivers of turnover in their organization, and to interpret those data through the lens of their organizational and environmental context.
•Carefully define turnover in terms of how voluntary, avoidable, dysfunctional, and pivotal each instance of turnover is for the organization.
•Develop a consensus for costing turnover, including flexible costing formulae by unit, skill set, job type, and location.
•Develop a consensus concerning retention priorities consistent with overall business strategy and economic conditions.
•Continually track turnover costs, rates, and functionality overall and by key categorizations.
•Develop systems that enable managers to track turnover metrics in their work units in real time.
•Study systematic reviews of turnover evidence and principles (such as this book!)
•Educate managers about decision biases and how to overcome them.
•Conduct external benchmarking to assess the extent to which turnover metrics place the organization at a competitive disadvantage.
•Conduct internal benchmarking to assess change and trends in turnover metrics over time.
•Conduct internal needs assessment to assess the impact of current and future organizational operational tactics and strategic plans on the supply and demand for human capital of particular types.
•Conduct external needs assessment to assess the impact of labor market conditions, economic conditions and trends, and industry trends on the supply and demand for human capital.
•Collect multiple types of data concerning the drivers of turnover in your organization, such as through exit interviews, post-exit surveys, current employee focus groups, survey linkage research, predictive survey studies, and qualitative research.
•Use well-developed survey measures with appropriate reliability and validity information available. (“Reliability” refers to a measure’s consistency across time, location, and circumstances. “Validity” refers to the extent to which a measure adequately captures the intended construct.) A useful resource, the Measures Toolchest, is maintained by the Research Methods Division of the Academy of Management: http://measures.kammeyer-uf.com/wiki/Main_Page
Final Thought
Employee turnover can be a surprisingly complex phenomenon. Organizations routinely develop extensive metrics, analytical tools, and dashboards to help managers make real-time strategic decisions about managing all types of resources—decisions about retaining talent warrant the same kind of attention. Fine-grained data collection, analysis, and consideration of context can help managers be more strategic about using evidence to develop comprehensive strategies for keeping the appropriate human resources in the organization.