The intersection between HR strategy and HR technology

Dynamical Systems

systematicHR Avatar

Forget for a moment that I’m talking about a mathematical theory.  At the core of this post is how we apply some quantitative reasoning to our ability to look at data and predict the future.  Personally, I don’t think we do enough to extend and create meaning from our data.  If we looked at data with an increasing view of the quantitative sciences, we would have a proportional increase in our ability to apply “art” to our interpretation.

We predict things intuitively every day.  Driving down the road, we watch other cars, measure their velocity, and predict if we are going to collide or not – and based on those predictions we often change our course.  In Human Resources, we should be predicting the direction and velocity that our workforce is travelling in a number of ways.  Not only do we watch the size of the workforce, but productivity, engagement, competency growth, and any number of other factors.

A dynamical system is a mathematical construct that predicts what the state of a particular object will be at a given point in time in the future.  Dynamical systems are useful in a limited way – through them we can predict future state outcomes in a limited number of variables.  But the point is that we can predict vector and velocity – in other words, if the workforce is travelling in the right direction and at the right speed.

First off, I should absolutely admit that HR is not comprised of a bunch of mathematicians.  This should be no surprise to anybody.  However, our organizations that make up the analytical arms of the HR organization, those who generate analytics and create meaning out of data, should have some ability to quantitatively view the data and understand trending, directionality, velocity, and the general parameters of the vector.  I hate taking us all back to high school calculus, but we remember the first derivative of the curve is the slope of the curve.  This is general trend of what our workforce metric is doing.  However, the second derivative of a curve is the determination of whether the curve’s direction is slowly shifting.  In other words, we might know that our overall turnover rate is dropping, and that is a positive sign.  However, do we know if the turnover rate is dropping more slowly than it was last quarter, even if the rate is better than it was before?

While most of HR is not in the mathematics field, and most of us don’t want to be anywhere near it, we should be applying some of these theories to our analytics.  If we look at any of our programs, we can see acceleration of desired outcomes after specific events, and deceleration of outcome after we have not done any change management or communications for a while.  We should be looking at competencies not as a growth metric, but as an acceleration curve over time.  Growth is good, don’t get me wrong, but acceleration is better.

I’d like to think that all HR data analysts out there are taking a serious look at the data they are presenting, trying to create and extend meaning beyond what is being requested, but today’s reality is quite mundane, even with all the cool business intelligence and dashboard technologies out there.

systematicHR Avatar

5 responses to “Dynamical Systems”

  1. Vince Lammas Avatar

    Hi Systematic,

    You are right to highlight the point organisations should take more notice of the trends observable in workforce data. The fact many HR practitioners (“people” people) are so averse to handling data does not help business in this respect.

    Too often the workforce stats are delivered in a way that answers the wrong kind of questions (i.e. What is the absence rate this month?) rather than the right kind (What do the short term and long term trends tell us about past and current workforce behaviours and how should we react?)

    This requires data supported with narrative and some useful “prediction”. Your point about the changes in the curve are key here I think – though I know many HR people who would stare blankly at your calculus reference.

    I also like your analogy with driving, but our brains cope with the everyday based on a dynamic model of the world which is tested and evolves each second.

    When observing oncoming traffic, I respond to “danger of collision” with a small adjustment to my left – in anticipation the other driver will do the same (we are clearly in England). It’s that shared construct that makes safety possible and regularly validating my world model.

    Do we have similarly reliable models of workforce behaviour (managers and employees0 that can be used to guide our actions when the data tells us something is happening?

  2. Andrew Marritt Avatar
    Andrew Marritt

    There are numerous reasons why such prediction is not well practiced. The biggest I guess is a disconnect between deep business understanding (inc building a mental model) and having analytic skills in HR. The other key one is quality and scope of data.

    A big issue with HR data is the relative low frequency of changes in the dataset. This is compounded by the fact that many organisations don’t really have a homogenous workforce. For a time-series analysis perspective this makes separating trends from noise difficult in a rigorous manner.

    As to scope of data, whilst HR often has large quantities of data it often is missing some key information. For a hiring / retention perspective I believe that external labour market confidence data is key though rarely is this built into workforce models.

    I would have hoped that techniques such as system dynamics and even agent based modelling would be of more use but realistically you have to provide so many caveats of look at a very short term that their use as providers of forecasts are limited. Understanding an underlying model can be used to provide understanding into likely causal relationships.

    I find my own work coming more and more towards data visualization. As an example, Only last week I created one report which showed learning by recruitment agencies plotted over time (effectively a scatter plot of key ratios with paths showing how the points change over time). Visualizations like this facilitate good conversations, in a much greater way than discussing rate changes where many find understanding difficult. Mapping HR data onto maps can also provide tremendous insight in certain circumstances.

    But I digress. What is increasingly clear for me is that the technology isn’t the problem (especially when there is so much good stuff like R and Rapid Miner which are free, and where tools like Tableau enable sophisticated analysis to be done quickly, without much training and at low-cost). I would put the three main issues as data quality / availability; ability to build sensible models mentally (linked to business knowledge); and confidence with basic data manipulation / statistics. When HR lacks one or more of these they tend to think a new software implementation is necessary.

  3. […] Dynamical Systems Systematic HR Monday, October 11, 2010 However, our organizations that make up the analytical arms of the HR organization, those who generate analytics and create meaning out of data, should have some ability to quantitatively view the data and understand trending, directionality, velocity, and the general parameters of the vector.  Copyright © 2010 systematicHR. READ MORE […]

  4. […] Read the full post from Systematic HR ReTweet […]

  5. […] I’m not arguing what set of technologies is more deserving of our approval in terms of market cap, but how these images inform our ability to interpret data in the absence of external influences.  The reason that I’m an advocate of business intelligence analysts that view data and operate in a function that is not touched by the externalities of the “real world” is that they can touch and feel the data with an objective eye.  Those of us who operate in business and business process can often be blinded by prior results that are not directly related to the data, but we think they are.  I’ve seen organizations completely disregard employee engagement surveys that identified terrible managers simply because productions or sales happened to be “hot.”  We’ve been influenced by circumstantial evidence that very senior managers can’t be “messed” with or that diversity data is better than it really is.  We’ve completely missed the mark on interpreting trend lines because we don’t have analysts who know how to look at data from a mathematical perspective. […]