Don’t Just Crunch the Numbers on Talent

Statistical analysis for talent management is the latest hot topic. Everyone seems to want to play, and many don’t know what they are talking about.

University HR programs are pushing this. Conference companies are rushing to market with offerings. Journals are carrying more than one article per issue on it. Of course, consultants who didn’t know how to spell it two years ago are suddenly leading sources for analytic methods.

Nevertheless, this is good news. Research has proven that path analysis techniques — including any form of multiple regression analysis, factor analysis, correlation analysis, discriminant analysis, or multivariate and covariance analyses — are powerful tools for finding root causes as well as predicting future value from current investments.

For years managers believed their experience was better than that of a bunch of number crunchers. Today, most managers accept that their estimations are seldom as accurate as a statistical analysis.

In case you have been on the moon or living in a biosphere for the past year, “quants” stands for quantitatively oriented people and programs. Quantitative or statistical analysis is not the same as strategic analysis, however. Statistics deals with numbers. Strategic deals with thought. Many people start by gathering some numbers they hope are related to an issue and then run some quant exercise to be sure at the end of the process they will have an objective result.

The potential problem is that their output is only partially relevant to the real issue. The amusing thing is they go along believing they really understand what is happening. Because they have not gotten to the root of the real problem, soon thereafter they have to run a do-over.

That premature number obsession is part of the “what should we measure” syndrome. If you are doing an after-the-fact evaluation of a previous intervention, that is a good question. But if you are talking about analysis prior to making the investment, the requirement is much different.

Analysis does not start with numbers. It starts with thinking and asking questions. It starts at the strategic level looking for the macro forces that affect the way you manage talent today, and more importantly, how you will manage it tomorrow. It is about constructing and understanding the context within which you now operate and will operate for the near future.

Case in point, a prospective client wanted to evaluate the ROI of HR services. As we talked, it was clear the company had not spent much, if any, time consciously aiming its services specifically at the corporate KPIs. Most say they do, but few companies can show direct connections.

Given that gap, what difference would it make which numbers we come up with? I had to take executives back to the beginning and ask questions such as:
• What are the major initiatives of your company now and into the near future?
• What outside forces impact your company?
• What is happening inside your company that is helping or hindering accomplishment of objectives?

Once we got clear on those types of strategic issues, the rest of the analysis went quickly. Going forward the company was able to design, predict, invest and evaluate all in one system.

Quantitative analysis is much easier than strategic or predictive analysis. Clarity is the first step. To hand a sharp strategic analysis picture to a quant, you will have to answer the aforementioned questions and several more like them. Then the quant can apply the appropriate path analysis technique to either predict a future outcome or evaluate a recent investment return. Measuring is much easier than determining what to measure.

If there is a trick to strategic analysis, it is to forget for the moment that you work in HR. Make believe you are a top executive in your company.

What problems and opportunities are your peers dealing with around marketing, sales, finance, production, customer demands and competition? Your services have to support those business issues. Once you see the connections, you can start down the path to identifying the things the quants should be evaluating.