Of all the places business intelligence can be applied, human beings are the most crucial. Most of the value of a modern company lies in intangibles such as expertise, loyalty, and relationships. Providing direction about what investments work, predicting and reducing turnover, and pinpointing engagement problems is what I do. Working with stakeholders to ask the right questions, picking the right key performance indicators, finding data, analyzing it, and reporting the results in a useful way, is what I do.

Our approach is to collect data from different sources and blend it.   That data might include engagement surveys, exit interview information, demographic information from human resources information systems, customer satisfaction data, and, perhaps most importantly, operational data about business metrics.   This approach is often not at all, or not done well.  I call this level synthesis statistics.   Viewing your data through this lens often provides novel insights into the running of your organization; too much management is done through myth and anecdotes, not through science.  You may not realize that your "mandatory" training program only reaches 60% of your employees, or that 80% of sales come from 20% of the representatives.  The next level is predictive analytics.    This tells you how your metrics inter-relate, and which investments are providing a positive return on investment.  It is predictive in that it provides a plan for allocating future resources, and makes predictions about what the return would be.

The final level is optimization.  Optimization has several key components.  

Segmentation is the most important .  Impact is broken down on a group-by-group basis to understand where the greatest and least impacts are achieved.  For example, a sales training program may be successful at boosting sales overall. However, segmentation might reveal that new sales representatives receive a great lift in their performance, while seasoned representatives show little to no benefit from the training. You know now to target new reps with training and find another intervention for the seasoned reps.  In some cases, the overall effect might be negative overall, but some groups may show a benefit.    

Mixture is the second key concept.  Most companies are spreading their human capital investments among a number of different pots on the human resources stove.   In fact, with content available from a variety of knowledge providers, and many customized programs, there’s potentially hundreds of separate investments, and the combinations can become astronomically large.   Which investment, or investments, to choose from?  Two important concepts emerge from the mixture optimization:  synergy, and diminishing returns.    Synergy is the idea that two things may combine to offer something more potent than the sum of the originals.    Examples abound:   John + Paul + George +  Ringo forming the Beatles; drugs that may have a very different and unintended effect when taken with certain foods and beverages; a coaching program allows the participants to implement the theories learned in their training program.    Diminishing returns are programs where multiple items are combined, and the net effect is less than the sum of the parts.  Diminishing returns are common as well, and for a variety of reasons.  One common issue is that there is only so much improvement to go around – “low hanging fruit” may be a common benefit taken by more than one program.   

Saturation is the third concept in optimization, the idea that the return on investment may not be a smooth, straightforward calculation.  Getting a return may require a certain threshold of investment to show any returns at all, or may drop off after a certain amount is invested.   There are anecdotes about training programs that require a certain penetration before the idea can catch hold.   People report returning to the workplace and finding out the new revolutionary ideas they have gotten are not welcomed.   They may not be specifically discouraged, just that there may not be support for the ideas “until the rush is over”.  The rush, of course, never slows down, or at least until the trainee has forgotten the new ideas, or gotten the message that the ideas do not belong in the practical work space.   

Metric Interaction is the fourth key concept.   Many analyses methods tend to focus on a single outcome.  This is necessary to an extent, to simplify the world to a version that can be analyzed.  Some of the tools, like regression and general linear models, focus on a single outcome variable in their mathematical models.   It is necessary, however, to consider how the different measurements combine and interact.  This is a positive attribute of the pragmatism of the ROI school-of-thought – it lends itself to combining and leveling variables in to their financial contributions.    An example of metric interaction is the known tradeoff in sales between units sold and profitability per unit.   There is an optimal point between the two.  It is possible to increase profit margin, but the effect on units sold will eventually appear.    

Time Line is the fifth concept in optimization.  Time lines require the way we sample the data to be broken down more than just “before the investment” and “after the investment”.   Investments may be said to have varying degrees of “stickiness” – their impact may be lost after a period of time.   Motivational components, in our experience, are much like this.   After some number of months, motivation may return to the baseline levels.    Another example of programs that become “less sticky” over time is technical training that focuses on aspects that are prone to change over time (for example, sales training that provides training based on the current year’s models; new equipment whose advantages fade as the industry standardizes to the new level of technology.)     The other side of “stickiness” is “lag time”.  Some programs may take some period of time to show their advantage.   Industries with long sales cycles provide good examples.   If a sales training program provides advantages in filling the “pipeline” for products that may take eighteen months between initial contact and closing the deal, measuring the effects after six months may not show anything.