Yes, We are All Individuals

Number 6: What do you want?
Number 2: We want information.
Number 6: Whose side are you on?
Number 2: That would be telling, we want information, information, information.
Number 6: You won’t get it.
Number 2: By hook or by crook, we will.
Number 6: Who are you?
Number 2: The new Number 2.
Number 6: Who is Number 1?
Number 2: You are Number 6.
Number 6: I am not a Number, I am a free man!

The Prisoner (1967-1968)

We know that part of being able to improve business results through talent is being able to measure talent in the first place. There are challenges not only in being able to do that, but also in the resistance to doing that. A recently published book, “How to Measure Anything: Finding the Value of Intangibles in Business” by Douglas W. Hubbard, covers three common objections to the measurement of intangibles. The first, and perhaps only truly valid one from a business perspective, is the obvious economic one where the cost of measurement exceeds the benefit. The second is based on the common misunderstanding of statistics that it can be easily manipulated (“Lies-damned lies-and statistics”), which can be addressed through better understanding of statistics, probabilities, and risk. The third, and perhaps most relevant to the topic of talent management, is the ethical objection, based on the perception that it is dehumanizing and threatening to measure certain things about people, such as their value to the company (current and future), their risk of loss, and other touchy subjects that are often brought up in HR predictive analytics.

Hubbard argues that decisions are still being made regardless and that making those decisions under intentional ignorance could be worse, so measurements need to still be made. Of course, the question is which measurements are really useful and pertinent to the decision. That at least gets the ethical discussion going in the right direction. In fact, it bolsters the argument that:

1. The decision should be connected to achieving business goals (i.e. the impact on success).
2. The decision in turn should be driving what measurements are needed.

If the measurements are pertinent to the decision being made and the decision helps achieve business goals, there is at least a logical purpose to the measurements and therefore the benefits can be weighed against the ethical issues. This is better than mindlessly gathering data for its own sake, especially if the data sheds unflattering light on the population being measured.

What gets people upset is if the measurements have an error factor (and almost all do), then a member of that population runs the risk of being described unfairly or worse yet, completely false. In addition, if there are correlations between the measurements made on a person and some undesirable outcome, condition, or risk factor for the population, then people are afraid of being pigeonholed, ejected, or otherwise unfairly labeled.

The problem is not in the measurements themselves, assuming they meet the criteria of being pertinent, but the way in which they are acted upon, which is usually management’s fault. For instance, a company might find that certain factors correlate highly with employee theft across a sample size of the company’s workforce. Used properly, that would help the company develop effective programs, policies, etc. that would lower the amount of employee theft. Used improperly, managers would target or otherwise treat employees differently who, while they might exhibit some or even all of these factors, never stole from the company. The point the managers missed was that the measurements and correlations were useful tools to see if the programs and policies were effective and that’s where it ends, period. They are not about predicting any one individual’s behavior, nor set them apart as “high risk”, etc. That’s the marvelous thing about people. They’re all different and managers have to remember that fact when they study population data. There is no shortcut in managing people.