4 Steps to Get Started with Workforce Planning Analytics

Today’s post comes to us from board member, David Creelman. When budget planning season rolls around, will your workforce planning analytics get you your fair share?

Workforce planning is one of the oldest areas of HR analytics. You can find many articles from the 1960’s on workforce planning—or as it was known then “manpower planning” (see for example Models and Modelling for Manpower Planning by W.R. Dill et al in Management Science Journal, 1966). The basic principles haven’t changed much in the last 50-odd years, what has changed is our access to data and the tools we have to manipulate it.

Despite its maturity, workforce planning analytics can be a frustrating topic to address. No matter how good our analytics tools, we are still making predictions about an uncertain future based on managers’ estimates about what the business will need. Also, managers may be unclear about what specifically they want from workforce planning analytics, leaving analytics pros with an unmanageably large task.

Here are four steps that will guide your approach to workforce planning.

Step 1: Get clarity of purpose

The most important step in workforce planning is getting clarity on what management will do with the analysis. For example, if they are interested in making a budget forecast over the next three years then that will require a different kind of detail than if the primary concern is forecasting the technical skills needed over to coincide with a planned product launch.

Managers may lean towards thinking that analysts can do a single forecast that covers every possible use of workforce planning but that’s unrealistic.

When you discuss their needs, make sure you cover these key factors:

  • Decisions / Actions. Learn what actions will be based on the analysis; these might include:
    • Budgeting decisions
    • Hiring plans
    • Decisions on the internal movement of staff
    • Training programs
  • Time frame.  Workforce planning typically looks ahead one to three years, however, it could be much longer, and, in some cases, it needs to be shorter.
    • If you are forecasting the need for unique skill sets that can only be developed in-house such as senior nuclear engineers, then you may well need a ten- or twenty-year plan—but not for all jobs, just for nuclear engineering ones.
    • If the issue is staying ahead of hiring needs for an hourly workforce, and filling a vacancy takes about a month, then one needs a fresh forecast of hiring needs each month—again, that’s not needed for all jobs, just the hourly ones.
  • Granularity.  For some purposes, broad headcounts will be enough, for others, a much more granular analysis will be needed.
    • Do you need to know how many salespeople you need city-by-city or is a national forecast sufficient?
    • Do you want to know overall managerial headcounts or function-by-function?
    • Are you more interested in skills rather than jobs (e.g. you want to have enough Python programmers, not just enough programmers)?
  • Accuracy. Discuss how accurate the analysis needs to be.

Step 2: The organization’s demand for talent

One half of workforce planning analytics is determining the demand for talent. Typically, this information comes from both:

  • Top-down forecast: Leadership explains their strategic plan (e.g. we will grow revenue by 10% per year) and what workforce size they see as acceptable.
  • Bottom-up requests: Each manager explaining what they need to meet the strategic plan.

This will have to be reconciled—and that reconciliation is more political than technical. The analytics pro should avoid getting caught in the middle.

You’ll need to have information about all the clarity of purpose factors to know how detailed the forecasted demand for talent needs to be.

In all cases, we are dealing with a combination of hope and guesswork. It often makes sense to run several scenarios to reflect that uncertainty.

Step 3: The organization’s supply of talent

Analytics pros are usually comfortable studying the supply side of talent. Here they can model turnover, retirements, lateral transfers, and promotions to predict where the current workforce is likely to be in various time periods.

Forecasting talent supply can be done in a rough way on the back of an envelope or by using sophisticated models informed by past data. Choose a level of analysis that fits the need and remember that the accuracy of the forecast supply of talent should be aligned with the accuracy of the forecast demand for talent.

Step 4: Filling the gaps

It’s self-evident that once we’ve forecast supply and demand we can determine the gaps. Once we have the gaps, we can take action to address them.

Determining how to fill the gaps opens many interesting and important opportunities for analysis:

  • What hiring processes will be needed?
  • Will the state of labor markets in the future make hiring harder or easier?
  • Can we fill gaps through training & development?
  • Will working with community colleges create the right talent supply in the future?
  • Can we reduce the size of the gaps through retention programs?

This kind of analysis draws on the creativity and insight of the workforce planning analyst. It’s here where we move from seeking to understand the future to recommending actions that will take the organization into the future. If this step is not executed well, then the earlier work will be in vain.

Workforce planning is a complex process and there are many decisions that need to be made before you start the analysis so that it produces accurate and actionable recommendations at the end.  The good news is that if you have the right tools and clean data then the analyst will be able to keep their eye on this end result, rather than be overwhelmed by the manual work of simply getting the analysis done on time. We could be entering a golden age of workforce planning, but only if we don’t underestimate what it takes to point the analysts in the right direction.

Please share your comments

This site uses Akismet to reduce spam. Learn how your comment data is processed.