As Federal contractors begin to utilize the 2010 EEO Census File to generate availability data, it provides a good opportunity to remind plan developers to take a look at how they are calculating their availability pools. Contractors need to avoid going on autopilot when it comes to defining their affirmative action plan structure. Things have changed and contractors need to visualize the impact those changes have on their goals and resulting outreach efforts. Plan developers should regularly undergo efforts, at least annually, to ensure that they are creating accurate, meaningful and defensible analyses. This is particularly important at a time when Good Faith Efforts made towards reaching Placement Goals are squarely in the spotlight during OFCCP audits.

Most affirmative action practitioners are now familiar with how OFCCP audits have (seemingly) become full-scale EEO investigations since the current administration moved in. While that is not the case in every audit, as I have heard of many that closed without incident, for the most part the new Active Case Enforcement directive has put contractors’ selection process front and center in all investigations. This increased scrutiny has further exposed the already vulnerable selection process with hiring getting most of the attention.

Since generating availabilities are the driving force behind the creation of Placement Goals, leading directly to Outreach efforts and expanding record keeping requirements, it makes sense to put in some extra effort and ensure the numbers reflect reality.

Let’s take a look at the key components of generating availabilities and consider the pros and cons of various decisions.

  1. Assigning jobs to occupational census codes – This step is often assumed to be pretty straightforward and yet it doesn’t receive enough attention. For starters, every job title should be assigned to an individual occupational code, and while some people choose to assign multiple codes to a single title, the concern is that not enough effort is being placed on the job to code connection. There are useful support materials available from the census bureau to help contractors match their jobs to detailed occupational codes including mapping tools to move from the 2000 EEO file to the 2010 file. One suggestion would be to develop a master list of jobs and their corresponding census codes, EEO category, job group and location. This will help the contractor review the data periodically to add or remove jobs as new positions are created or the scope has changed. Also, job descriptions and input from department managers can be useful here.
  2. Creating job groups – It is surprisingly easy to create flawed job groups because people are in a hurry to lump jobs together to generate a required report. In today’s compliance environment, there is a lot of motivation to make sure job groups are well organized because not only are placement goals based on them, but the same groups are used in adverse impact and compensation analyses. It isn’t fun, or wise, to use job groups in a plan only to try and redefine them later when problem areas show up during an audit. Contractors often develop job groups starting with EEO-categories, then break them down further based on the recruiting practices, pay structures and movement of jobs. For example, the Professional category might be separated out into Professionals A-Engineers and Professionals B-All Others with different jobs being linked to each. Technically, if a plan has fewer than 150 employees, contractors can stick with EEO categories for the job groups, but any size company should be more interested in their specific environment than just defaulting to a pre-defined and generic grouping of jobs. A few other tidbits when creating job groups would be to avoid unnecessarily large numbers of employees in a single group where possible. If there are hundreds of people in a single job, there isn’t much you can do, but if you have a job group that has many jobs in it including a single job with a disproportionate number of people in it, then you might be drowning out the other jobs, creating a skewed availability pool. On the other side of the coin, having too many job groups with a small number of jobs in them might reduce a contractor’s ability to generate any meaningful goals, not to mention the fact that it could look like you are disguising an issue through the “divide and hide” method. The bottom line is that job groups warrant a review to ensure that they make sense.
  3. Creating external availabilities – Another challenging element of generating availability pools is the identification of external availability pools. While the regulations tell contractors to use the most current and relevant data, this leaves a lot of latitude for contractors to apply some art to the science of affirmative action planning. The point is for the contractor to put a little effort into aligning the external data, typically EEO census data, to the job group and not just defaulting to the same labor area for every job group. Some jobs/job groups have a local recruitment focus while others clearly focus on broader areas such as statewide and nationwide recruiting efforts, not to mention groups that recruit from certain schools or associations. It is also not unreasonable for the total external pool to reflect more than one category of external data and be weighted and merged into a single external pool. If there are targeted local recruiting efforts and separate, targeted efforts made across a state, region or the country, then there may be value in applying a structured mathematical formula to account for it.
  4. Creating internal availabilities – Selecting internal availability pools is often as simple as looking at the job group(s) within the AAP who are eligible to be promoted into the group in question, using them as the comparison group. However, other methods can be considered such as; assigning training programs that made employees eligible for promotion, linking individual jobs within the establishment, or perhaps utilizing jobs and job groups from other establishments within the company. Note that an argument can be made that linking job groups from other establishments might be considered external recruiting.

    Contractors should note that there are multiple, reasonable methods for determining the availability pools but there are a few trap-doors as well. Let’s address some of those concerns in the next section.

  5. Weighting of the two factors – When it comes to finalizing availabilities, contractors must group the external and internal factors together and weight them based on expected recruitment patterns. The intention is to generate a recruitment data pool that reflects the candidates available for hire. Note that the regulations provide wide latitude here and OFCCP, consultants, attorneys and many Federal contractors have their own unique opinions on how to generate those pools. There are several ways to identify the factors and weight them to develop goals and each has its pros and cons.

    Using historical transactions to drive weighting – My perception is that using recent hiring and promotion data to drive the factor weights is the norm in affirmative action planning, and why shouldn’t it be? History shows the actual selection pattern and it reflects reality, right? It doesn’t hurt that automation can take the data and spit out the results in a few seconds, “no muss, no fuss”. Not only that, but I have seen OFCCP expect it in an audit and everyone wants to appease OFCCP in an audit, right? Well, there is potential for a backlash here as well.

    • Just because the history of recruitment identifies what has happened in the recent past, that doesn’t mean that company record keeping is lock step with it. How clean and accurate is the data? Contractors often struggle to track the details and any gaps here will likely have a big impact on the analyses.
    • Does the recent history align with the current process? Often companies have mergers, job realignment, or other unpredictable activities such as a significant increase or decrease in recruitment. I have seen hiring and promotion vary greatly from year to year which could provide the contractor with data anomalies in the history data which could distort the results. I have even seen demotions marked as promotions further complicating availability pools.
    • Beware of fluctuations between internal and external weighting. If history provides that one year the selections were primarily external and the following year they were mostly internal and the activity is “light”, your goals could swing dramatically which may not pass for a realistic view of who was actually available for selection.
    • Sample size matters – If the volume of selections is large, it is reasonable to speculate that the results would be balanced, but what about the opposite scenario? If you have a high volume of external hires followed by a period of slow hiring where there were much fewer numbers, then the resulting analysis could skew. While people might assume the results would balance out across multiple years of data, the question remains, “are you sure?”

    An alternate method for weighting availability factors might be to use a pool of what is commonly known as SME’s or Subject Matter Experts. If a contractor has experienced staff such as recruiters and hiring managers, it is reasonable to believe that those employees could look at the various job groups and identify who the recruiting pools are and at what frequency they are feeding the selection process. Similar to the use of historical data, this option makes sense and is commonly used as a preference to or in place of automated data functions. Once again, there are a few shortcomings to acknowledge.

    • Experience needed – Does the company have the experienced team to identify the selection patterns? Without experienced team members, it would be difficult to conduct a manual review.
    • “Guesstimation” can be dangerous – By using opinions, you are banking on accurate knowledge and this is not necessarily as automatic as hard data. Then again, it’s worthy of debate because the manual review might be more reliable than the data that is available, especially if there is a lack of depth within the data tables.

    The method I would prefer to use is probably obvious by now, and that is to employ both automated and manual if possible. If a plan developer has the ability to manipulate historical data (using multiple time periods for a larger sample is recommended) to drive some initial results and then the SME’s can take time to review the information and remove any anomalies, then it would appear that the result would more likely survive scrutiny and provide meaningful and defense-able results.

Conclusion – The purpose of this exercise would be to help those folks out in the field recognize the strengths and weaknesses of each method for calculating availability and continue to work towards a result that converts to recruitment efforts that reflect the legitimate needs of the company as opposed to leading them off course resulting in greater gaps that could be exposed during an audit.

Good luck.