Applying Quantitative Staffing Analysis to Succession Planning
There are several ways in which a quantitative staffing analysis can enhance the effectiveness of an organization’s succession and development planning process. To begin with, quantitative analysis can help an organization identify specific areas of vulnerability and determine the “right” size of the candidate pool. Each of these two applications is discussed later in the chapter.
Note that for all of these analyses, your data should be segmented by whatever job categories you are using in your succession analysis. For example, if you are applying succession and development planning at specific job levels (e.g., all vice president– or director-level positions), then the analyses I am about to describe should be segmented by job level. If your succession planning process focuses on specific job titles or positions (e.g., Vice President—finance, General Counsel), then these analyses should be conducted for each job separately.
Here are two examples of how workforce planning methods can be used to strengthen succession and development planning.
Identifying Vulnerabilities and Forecasting Losses
Through quantitative staffing analysis, your organization can analyze management demographics and identify the impact that losses might have on the number of managers and candidates that will be available at various points in the future. Usually, these losses are due to retirements and voluntary turnover. An understanding of what losses might occur when can help you to fine-tune your succession and development plans and ensure that an appropriate number of candidates is being developed (not too many or too few).
Quantitative analyses can strengthen your process in two ways. First, these analyses can help define the number of openings that can be expected in the coming years. This can help you to refine your estimates of the number of candidates that are needed and when candidates with those capabilities should be “ready.” Second, they can help to identify those individuals who are most likely to leave (and thus the positions that may require more in-depth analysis).
As you structure this workforce planning–type gap analysis, make sure that you use an appropriate time frame. It is often helpful to consider planning periods that coincide with your estimates of readiness. For example, the first planning period might be from “now” (i.e., the current date) to a point three years in the future (to address “ready in one to three years” candidates); the second planning period might be from three years from the current date to five years out (to capture those candidates that might be ready in three to five years). If you need more details, a more complete description of these retirement analyses is included in an earlier chapter. Here are two areas to focus on.
Forecast Retirements
Retirement analyses often play critical roles in identifying areas of vulnerability. Clearly, very specific succession plans can and should be produced whenever an organization identifies specific individuals who are actually going to retire. What can be equally valuable is a more general analysis that looks at various retirement scenarios on a unit-by-unit or level-by-level basis. Once an organization has gained an overall sense of the retirement losses it can expect, it can make a more realistic determination of the size of the successor pool that it should be developing. This type of analysis is a “tailor-made” application of quantitative workforce planning, using the various tools and techniques described elsewhere in this book.
Some companies base their retirement projections on historical trends. This method is at best unreliable; it applies only when retirements are spread out evenly over a period of several years. In most companies, this even spread just does not occur. Projections should be based on an analysis that considers actual retirement information for the management pool being examined. Here are two methods that can be used to define the expected number of retirees. These methodologies are described in more detail in a previous chapter, but it is appropriate to review them here.
Average retirement age. The easiest way to consider retirements is to base your projections on the actual average retirement age. Calculate the average age at which managers in the group you are studying have actually retired. Where possible, consider retirements over the past four to five years when calculating these averages. Remember to exclude from your calculation any individuals who elected to retire early as a result of any “special offers” designed to reduce staffing levels. Try to consider only those individuals who retired “normally.” In some cases, average age could be calculated for a large group as a whole (e.g., all managers at or above a given organization level). In other cases, separate averages should be calculated to reflect specific conditions for certain subsets of your population. For example, you may find that managers in difficult overseas assignments or jobs that require extensive travel actually retire sooner than those who work domestically and travel less often do.
Once you have calculated your averages, simply look at your current population and note their ages. For planning purposes, assume that any individual who reaches the appropriate average retirement age will actually leave. Remember to count the number of individuals who will reach the average age during each planning period. Suppose, for example, that your average retirement age for a particular management group is 59 years old. Applying this approach, you would then assume that all managers who are currently 58 or older will retire in year 1 (when they would reach the average age of 59), all who are 57 currently will retire in year 2, all who are now 56 will retire in year 3, and so on.
Needless to say, this assumption may not apply to any specific individual, but it will apply when you look at a group of individuals. After all, that is the basis of an average—some managers may choose to work past the average age, but just as many will choose to leave before they reach that average age.
Also, make sure that you allow for different average retirement ages for different positions. Here are two obvious examples. Linemen who brave winter storms to make needed repairs to downed electric lines will retire at an earlier age than engineers at the same level who work primarily in indoor offices. People in high-stress positions (such as those in law enforcement and air traffic control) will retire at an earlier age than people in less stressful jobs. These differences can prove to be significant.
While simple, the average age method typically yields surprisingly accurate results. For many organizations, this method provides results that are absolutely adequate to support their planning.
A retirement “distribution.” In some cases, you may need to analyze retirements using a more sophisticated methodology than simply calculating and applying an average retirement age. One really effective way of estimating retirements is to assume that employees will retire over the course of several years, beginning with their initial year of eligibility.
To use this method, you must estimate or assume two things:
The number of years over which the retirements are to be spread. Determine the number of years over which retirements will be spread (e.g., for planning purposes, you will assume that most employees will retire within seven years of their first eligibility date).
The percentage of those eligible that will retire in each of those years. Estimate the number of employees (on a percentage basis) that will retire in each of the years in the distribution as defined in the previous step (e.g., of all the employees who become eligible in any given year, 20 percent will retire in the first year of eligibility, 15 percent in the second year, 10 percent in the third year, and so on). These percentages could be based on actual historical data or on estimated probabilities.
To apply this method to a particular group or job category, first create a list of the managers in the unit(s) that you are analyzing. For each manager, determine their retirement eligibility date by applying the normal criteria (e.g., age plus length of service needs to sum to 80). Print the list, sorting first by job level and then by eligibility date (in ascending order). Now count the number of employees in each category or level that will become eligible to retire in each year. Calculate the expected retirements by multiplying this number of eligible employees by the appropriate percentage for each year and summing across all years. Here is a simple example.
Suppose that we expect most employees to retire within three years of becoming eligible. More specifically, we expect that 50 percent will retire as soon as they are eligible, 20 percent will retire the next year, and 30 percent will retire in the third year. Our sorted list shows that 100 employees will become eligible in year 1, 80 in year 2, and 70 in year 3. How many retirements can we expect in each year?
In year 1, we would forecast that 50 people will retire (i.e., 50 percent of the 100 that became eligible in year 1). In year 2, we expect that 70 employees will retire (i.e., 30 percent of the 100 that became eligible in the previous year plus 50 percent of the 80 that are eligible for the first time in year 2). In year 3, we would expect 79 employees to retire (i.e., 20 percent of the 100 that became eligible in year 1 plus 30 percent of the 80 that became eligible in year 2 plus 50 percent of the 70 that are eligible for the first time in year 3). These calculations are easily handled in a spreadsheet.
Obviously, this is a much more complex algorithm than simply using average age, but it could yield better results. Make sure that the extra precision you obtain justifies the added effort you must expend.
As a final option, you might want to consider talking with the managers of those individuals that you have identified as eligible for retirement. Often, these managers have good insights regarding what these people’s actual retirement plans might be (e.g., “John will probably be leaving in the next year or two, but I think that Mary will continue to work long past her retirement eligibility date”).
Estimate Voluntary Turnover
In middle-level management positions and below, most openings are created not through retirement, but through voluntary turnover. Consequently, your analysis of vulnerability should include some estimate of the number of managers that will choose to leave your organization (e.g., to take jobs elsewhere).
Unlike a retirement analysis, which tries in some way to identify specific individuals who may leave, an analysis of voluntary turnover should focus on a group or “cohort” of managers, not on any managers in particular. Generally, voluntary turnover is estimated by multiplying an assumed turnover rate by the number of managers in that particular group. For example, if you think that the voluntary turnover rate will be 6 percent per year and there are 50 managers in the group you are looking at, then you might assume that 3 managers will leave during that 12-month period (i.e., 6 percent of the 50). It is important to remember that you are not identifying which 3 managers will leave—just that it is probable that 3 will leave (and may need to be replaced through succession).
When estimating voluntary turnover, try to group managers who have similar characteristics (e.g., look at managers by organization level).
When estimating voluntary turnover, don’t just blindly project historical rates. Estimate instead what you think turnover will be for the period in question. It may be helpful to start out by defining a historical rate, but then modify that rate as necessary to reflect the business conditions and economic climate within which you will be operating in the future. Here are two common situations in which historical rates must be adjusted:
Changing economic conditions. If the economy has been strong over the last several years but is now weakening, it is likely that turnover rates will go down. Conversely, when economic conditions are good and opportunities abound, managers are more likely to leave to accept positions elsewhere.
Changing competitive conditions. Needless to say, competition affects turnover. If competition for the talent you need is increasing (e.g., where a competitor is trying to grow significantly and is locating a facility in your area) then turnover will increase. If, however, your industry is retrenching, then turnover is likely to go down.
In either of these cases, it would not make sense to assume that the turnover rates for the future would be the same as those that were seen in the last few years.
Run a Gap Analysis Model
Once you have determined what your assumptions regarding turnover will be, run a modified version of the workforce planning gap analysis to determine the number of possible replacements that might be required and determine the number of candidates that might be needed to provide adequate depth of coverage for the positions you are analyzing. Identify additional candidates where necessary to increase pool size and create development plans for these individuals. Remember that you may want to be developing more than one candidate for each gap that you identify (e.g., your target might be two candidates for each opening). Figure 19-1 diagrams what this process might look like.
Figure 19-1. Using a Workforce Planning Approach to Estimate Losses Within a Succession and Development Process.
Maintaining Candidate Pool Size
Organizations often wrestle with the issue of how large their succession pool should be. For most, this problem is especially difficult because the company really does not know how many management openings it can expect during the planning period.
A quantitative staffing model can be quite helpful in defining the size of the candidate pool that is required in order to support your succession planning objectives. A gap analysis is needed, but it is built in a somewhat different way from the gap analyses that are usually included in a workforce planning process. A well-defined workforce planning–style model can help an organization to better understand the number of positions that are expected to become available at each level at various points in the future. Once these numbers are determined, proper pool size/depth can be defined (e.g., by applying standard ratios such as two candidates per opening).
Most staffing models (including the ones that apply here) include components that allow an organization to define future staffing requirements, identify current staffing levels, forecast future staff availability, and calculate specific differences between requirements and availability. Once these gaps and surpluses are known, the organization can then define the specific staffing plans and actions that are needed to align supply and demand. Usually, however, these processes include all positions and incumbents. Here, the same process is used, but it will be applied only to those individuals who are considered to be candidates.
To create a staffing model that defines candidate pool size and depth, follow these steps:
- Define staffing model parameters. First, define the time frame of your model. As stated earlier, this is usually consistent with the estimates of readiness that you use in your succession and development process (e.g., a period of one to three years and a second period of three to five years). Next, identify the population to be analyzed (e.g., the positions to be included in the succession plan and the positions from which candidates for those jobs might come). Finally, create a model structure in terms of columns (e.g., organization units or functions) and rows (e.g., organization or management levels, such as vice president, director, manager, and supervisor).
- Define actual staffing requirements. In this instance, staffing requirements will reflect the desired size and depth of the candidate pool you have determined that you need. Once the actual number of management positions at each level is defined, you will be able to determine quite easily the size of the candidate pool that is needed at each of those levels (and at each state of readiness, if you take that approach). Create and apply ratios of candidates to positions (e.g., 1½ or 2 candidates for each position or 1 candidate at each state of readiness for each position). This analysis will tell you how many candidates will be needed for each time frame (and thus how “deep” your candidate pool needs to be). Simply multiply the appropriate ratio by the number of positions at a given level.
It is the result of this calculation of the number of candidates that defines “demand” (in workforce planning terms, anyway), not the actual number of management positions. Thus, the subsequent gap analysis will define the staffing actions needed to maintain that particular candidate pool size—not simply what needs to be done to fill the openings that are anticipated. - Identify current staffing levels. Define the number of candidates (not necessarily incumbents) that are currently in each cell of your model. This will include the number of candidates currently being developed for each position or level, at each state of readiness. This number should reflect “what is,” not “what should be.”
- Forecast staff availability. Next, you will need to estimate the number of candidates that you will have at each level in each category of your model at the end of each of your planning periods. Begin with current staffing levels (as defined in the previous step). Next, estimate expected losses in each category as a result of retirement and voluntary turnover (perhaps using one of the algorithms described earlier). You may want to modify overall turnover and retirement rates to reflect what actually happens among candidates (e.g., candidates are more or less likely to leave voluntarily than managers as a whole). For now, do not assume any replacements; let the model tell you how many replacements will be required.
- Calculate staffing gaps and surpluses. Next, define staffing gaps (where demand exceeds supply) and surpluses (where supply exceeds demand) at each level in your model. Compare estimated requirements (defined by the desired pool size, not the number of management positions) to the forecasted supply of candidates in each category and identify where the pool may be insufficient. Generally, focus on those levels where there are not enough candidates. Don’t worry about instances where the pool at a given level may be deeper than is required.
- Define required staffing actions. Once specific gaps and surpluses are calculated, define the staffing actions that would “best” eliminate these differences and maintain the proper management candidate pool size. Identify what promotions and other job moves will be needed in each year to maintain that pool depth and ensure management continuity. Remember to identify any development that must take place in support of these staffing actions.
Keep your staffing strategies in mind as you define staffing actions (e.g., that to continue to generate a flow of new ideas, you intend to fill 75 percent of the anticipated openings from within and the remaining 25 percent through outside recruiting). This component of the analysis will also allow you to identify and address “cascade effects” (e.g., where promoting a candidate to fill a job at one level creates a new opening that must be filled by promoting a different candidate into the vacancy that was just created).
Figure 19-2 depicts what this version of a gap analysis might look like.
Figure 19-2. Using a Workforce Planning Approach to Maintain an Appropriate Candidate Pool Size.
Here is a simple example of what this approach might look like for one planning period (e.g., from “now” to the end of the third year). Suppose that your organization has three levels included in your succession plan (e.g., executive, director, and manager). Suppose further that your succession planning process incorporates three levels of readiness (e.g., now, one to three years, and three to five years). For the purposes of this example, let’s assume that you are evaluating the candidate pool size for all positions in one business unit (you could just as easily do this analysis for all positions across business units). Further, let’s say that your company has set the following ratios or targets for the desired number of candidates for each position at each state of readiness:
Two “ready in one to three years” candidates for each position
One “ready in three to five years” candidate for each position
In an actual case, there might be different ratios for different jobs and/or levels, but for this example, we will assume that the same set of ratios applies to all levels. Here is a step-by-step example of how this analysis might be structured:
- Define current headcount. Identify the current number of candidates that are actually being developed for each position at each state of readiness. These numbers should reflect “what is,” not “what should be.” Your numbers might look like this:
Current Number of Candidates Executive Director Manager Ready now 8 20 60 Ready in 1–3 years 6 34 80 Ready in 3–5 years 5 21 50 - Estimate voluntary turnover: Determine the rate of voluntary turnover you expect among the candidates. This will probably be a different rate from that for managers in general. Let’s assume that turnover among succession candidates is 5 percent per year. The rate that you would use is then 15 percent (5 percent per year for each of the three years in your planning period). Again, in a real case, rates might vary by level, business unit, or state of readiness. Applying that 15 percent rate to the current number of candidates in the example would yield the following turnover losses:
Voluntary Turnover Among Candidates Executive Director Manager Ready now 1 3 9 Ready in 1–3 years 1 5 12 Ready in 3–5 years 1 3 8 - Forecast retirements. Next, estimate the total number of retirements you expect to occur from the candidate pool across all three years of your planning period. (Other chapters of this book describe several approaches that you might use.) For this example, let’s assume the following retirements:
Retirements from the Current Candidate Pool Executive Director Manager Ready now 1 Ready in 1–3 years 2 1 Ready in 3–5 years - Forecast future candidate pool size. Next, you will need to forecast what you think the available number of candidates will be at the end of your three-year planning period, given your starting population and the assumptions you made about voluntary turnover and retirements. To get your forecasts, simply subtract voluntary turnover and retirement losses from the starting number of candidates. For the Executive/Ready Now category, six candidates will be available (eight initial candidates, less one voluntary loss and one retirement). For this example, the result would be:
Available Number of Candidates Executive Director Manager Ready now 6 17 51 Ready in 1–3 years 5 27 67 Ready in 3–5 years 4 18 42 - Define required candidate pool size. Here you apply your chosen ratios. We have assumed that in this example there are 8 executive positions, 22 director positions, and 64 management positions. Now, multiply your chosen ratio (i.e., 1 ready now candidate, 2 ready in one to three years candidates, and 1 ready in three to five years candidate) by the number of positions (not candidates!) to determine how many candidates are needed in each position at each state of readiness. For example, 44 ready in one to three years manager-level candidates will be needed (2 such candidates for each of the 22 management-level positions). When you complete those calculations, you get the following requirements:
Number of Candidates Required Executive Director Manager Ready now 8 22 64 Ready in 1–3 years 16 44 128 Ready in 3–5 years 8 22 64 - Calculate gaps and surpluses. Finally, compare the number of candidates that will be available to the number that will be required. For example, the model forecasted that 6 executive-level candidates in the ready now category will be available, and 8 will be needed. The gap between these is -2 (always subtract demand from supply to depict a gap as a negative number). Completing this calculation for the rest of the model yields this result:
Candidate Gaps/Surpluses Executive Director Manager Ready now –2 –5 –13 Ready in 1–3 years –11 –17 –61 Ready in 3–5 years –4 –4 –22 - Define staffing strategies and plans. Determine what staffing and development actions need to be taken to eliminate the gaps and surpluses that were calculated (thus achieving the desired candidate pool size). This might include identifying and developing additional candidates, recruiting additional talent that has the skills to be successors, identifying individuals working in other companies that could step in, and other such actions. Remember to keep the length of your planning period in mind when creating these plans so that your plans are realistic (e.g., would it really be possible to develop a new candidate in the two years that are available?). Remember too that if part of your solution is to elevate a candidate (e.g., from the three to five years category to the one to three years category), it will be necessary to replace that person with a new candidate in the category she is moving from (i.e., three to five years in this simple example).