Tuesday, August 11, 2015

A Mathematical Approach to HR?! : Promoting Responsible Use of Numbers in People Management

"If we must apply a Mathematical approach to HR, let us go beyond Arithmetic. Human Resource Management is more like a differential equation that can have multiple solutions!” I heard myself telling the Senior HR Leader. This was my fourth ‘encounter’ with this gentleman (See 'Passion for work and anasakti ‘, 'Appropriate metaphors for organizational commitment ‘ and ‘To name or not to name, that is the question’ for the outcomes of my previous interactions with him). This time, we were discussing the point of view that HR can get the elusive ‘seat at the table’ by being more data-driven, quantitative, objective and mathematical in its approach. Similar to what had happened during my previous encounters with him, this interaction also prompted me to think a bit more deeply about the underlying issues.

Mathematics and logic are immensely useful tools. The five and a half years of engineering education that I have received, my pre-MBA job as an Aerospace Engineer, the Social Research Methods related studies during my MBA and the initial years of my HR career spent in Compensation Consulting & Research Based HR Products  have made me very comfortable with quantitative and mathematical approaches to diagnosis and solution design.  Having practiced Six Sigma, I have experience in the process improvement approach of converting a physical problem into a mathematical problem, finding a mathematical solution and converting the mathematical solution into a physical solution.  I have also used quite a bit of statistics especially during the best practice & benchmarking studies during my five years in HR consulting. However, there was still something that was bothering me regarding my conversation with the Senior HR Leader.

Once I stayed with that feeling of discomfort for a while, things began to crystallize in my mind. The first thing that came to mind was an incident that happened a few years ago when my son was about 5 years old. I had bought him his first calculator and he was very excited.  For the next couple of days he was chasing me saying that “Tell me all your problems; I can solve them”.  It was an interesting task to convince him that most of the problems can’t be expressed in numerical terms and that even those problems that can be expressed in numerical terms can’t always be solved using the functions available in the calculator! When I thought about the matter a bit more, many other aspects came tumbling out:
  • Perfect Logic coupled with Questionable Assumptions : Logic is a great tool for reasoning. The problem is just that any system of logic is only as good as its assumptions. Great logical reasoning skills with wrong assumptions will just lead to a wrong inference faster. While this would hold good for any field, the risk is higher in HR, as the domain has quite a few unsubstantiated assumptions. Yes, over the last couple of decades a significant amount of research has been done in the Human Resources Management (HRM) domain. But the very nature of the domain imposes severe limitations on validating the assumptions underlying  HR related decision-making (See  ‘Research and a three-year old’ and ‘Truths stretched too far’ for a more detailed discussion)
  • Lost in conversion : When we look at applying the Six Sigma approach mentioned above (physical problem - mathematical problem - mathematical solution - physical solution) to HR, the difficulty is in ensuring that while converting the physical problem to mathematical problem the essence of the matter is not lost. Otherwise we might end up solving the quantifiable but peripheral aspects of the problem while the core of the problem (which is difficult to quantify) goes unattended. We must remember that many of the things that really count can’t be counted!
  • Misuse of Mathematical Induction: This occurs when one tries to apply a purely mathematical type of reasoning to a human process where it doesn't apply. I came across an excellent example of this in a HR Shared Service Centre (HRSSC). The Head of this HRSSC was a firm believer of setting 100% accuracy (zero error) as the performance target. His strategy for making this happen included a motivational talk to the employees with the following line of reasoning: "Can't you do one transaction without error? If you can do that what prevents you from repeating the same 12,000 times? This is all that is needed to make an 'error-free' year and meet your performance target”.  While the above approach seemed to be perfectly logical it was completely unrealistic from a performance management point of view. The transactions involved a large amount of manual intervention making it highly error prone. The ‘zero-error target' ended up de-motivating the employees (instead of motivating them) as they were highly unlikely to achieve it. This brings to mind a Zen Proverb - "Water that is too pure has no fish"!
  • Chasing the numbers: A related problem, that comes up especially when we try to quantify (because quantification is required for further processing) things that are difficult to quantify is that of making simplistic or overly optimistic assumptions to enable quantification and even to get the numbers that we want to get. For example, when we try to calculate the time required for doing a particular non-mechanical task (as the first step in estimating the required staffing levels), we often don’t take into account ‘invisible work’.  The invisible work arises from factors like complexity of the situation (that can’t be quantified easily) and the difference between ‘the process map’ and ‘the way things actually get done’. While in the case of the latter, it can be argued that the solution is to fix the process, it might be difficult in a situation when complex interfacing/influencing is required to do the task or in a situation where fixing the process is difficult at the level of the jobholder (as it involves fixing the ‘ecosystem’ around the process in addition to the process). Emerson was not too far off the mark when he said "The results of life are uncalculated and uncalculable. The years teach much which days never know"!
  • Banning complexity and complex motivations : Another problem comes out of a definition of rationality (a mental model) that is too narrow. As Mencken says, "to every complex problem there is an answer that is clear, simple and wrong"! In many areas related to people management, there are deep psychological factors operating that render purely ‘logical’ approaches ineffective.  See ‘Performance ratings and the above average effect’ for an example.  Similarly, when we consider only the (visible) employment contract and ignore the (invisible) psychological contract another set of problems arise (See the series on ‘Salary negotiations and psychological contract’ for more details). Another example could be viewing the interactions with the labour unions (say in the context of arriving at a Long Term Settlement) as a purely economic negotiation exercise. The reality here is that a union is a political entity with a constituency to satisfy. Hence, even if the management offers a ‘competitive deal’ (by industry standards), the union leaders might have (internal) compulsions not to accept it and resort to various pressure tactics (including demonstrations and stoppage of work) - just to convince their constituency (members) that they have done all they can to force the management to offer a better deal (or the best possible deal). 
  • House built on sand : We also have the interesting problem of processing/computing data without paying adequate attention to the ‘level of measurement’ that generated the data.  Typical problems involve taking ‘ordinal’ or ‘interval’ data and apply computing methods that are valid only for ‘ratio’ level data. This could be more of a problem in HR, since many of the HR professionals are not well-versed in quantitative methods. The numbers can give us a false sense of surety and doing Arithmetic operations with those numbers to derive inferences can give us a false sense of confidence on the decisions based on those inferences.  There is a huge difference between being able to calculate something and being able to understand it. If our objective is to influence that 'something', being able to calculate it without being able to understand it can create more harm than good. It often becomes very difficult to convince HR leaders who are ‘too sure of their numbers and calculations’ that HR process maturity takes time or even that ‘It takes 9 months to make a baby regardless of how many couples you put on the job’. This becomes very pertinent especially in those situations where a business leader or the CFO (without any HR background) has been moved into the HR Head role! This brings us to a more fundamental issue. The over-reliance on numbers sometimes indicates a (stated or unstated)  shift in the underlying paradigm for people management in the organization- from a relational paradigm to a transactional one. This is something that we must watch out for (See 'Towards a Philosophy of HR' for more details).   
  • Wishing away the paradoxes and dilemmas : People Management, by its very nature, is a field that is full of paradoxes.  A paradox occurs when there are multiple perspectives/opinions (doxa) that exist alongside (para)- each of which is true - but they appear to contradict/to be in conflict with one another. A paradox can’t be resolved in the same way a problem can be solved. To effectively deal with a paradox, we must wrestle with it till we reach a level of understanding (or wisdom) that enables us to see the paradox in a new light and arrive at the most appropriate solution in that particular context. Often, there are multiple solutions -making HR more like a differential equation (that has multiple solutions) and not like Arithmetic (where there is one right answer)! It can also be argued that dealing with some of the issues HR is even more complex than dealing with differential equations because in some of those paradoxical situations, the choosing from multiple ‘correct’ solutions is a matter of Aesthetics and not Logic! One can develop a keen sense of this 'Aesthetics' only through years of struggle with the paradoxes and dilemmas  in HR(See 'Truth and Beauty : Elegance and Motivations in HR' for more details)

So where does this leave us?  To me, best approach is that of ‘triangulation’, that combines qualitative methods with quantitative methods to develop a more comprehensive understanding of the reality. We should make an effort to figure out if the particular HR issue that we are dealing with is more like a ‘problem’ or more like a ‘paradox’ and deal with it accordingly (See ‘Making problems disappear’ for details). Data and analysis are very useful. But they are not substitutes for understanding and wisdom. Even when it comes to the matter of strategy making, it has been argued that the core strategy making process is essentially intuitive, with data & analysis being useful as an input/trigger for strategy making and also as tool for doing a reality check on the strategy created.  The same holds in the case of HR strategy also! We must also remember that in the physical world (outside Mathematics) there are 'singularities' where 'normal rules/algorithms' no longer work!

Similarly, benchmarking is definitely a very useful tool. But benchmarking should be done with the context also included (and not just the numbers).  For example, benchmarking a ratio like the ‘ratio of the total number of employees to the number of employees in HR’  can be misleading without the understanding of context specific factors like the mandate/deliverables of the HR function, the HR operating model, the degree of outsourcing, the degree of automation (degree of Employee and Manager Self-Service), profile of the workforce etc. Casual benchmarking, like casual sex, is easy but dangerous!  We must also ensure that HR processes and practices follow from the HR Philosophy of the company and not the other way around (a common problem that arises from the obsession with 'best practice benchmarking'). Yes, we must leverage numbers and the power of numbers in HR. However, let’s use them responsibly - by ensuring that the numbers and the calculations accurately reflect the underlying reality!

Any thoughts/ideas on promoting more responsible use of numbers in people management?


Unknown said...

I read this with interest. The balance that you call for can still be achieved when thematic contextual complexities are quantified through mathematical processes. This implies that the mathematical equation becomes a representation of the contextual complexities. The implication then is that any mathematical result then represent a certain contextual complexity. If mathematical principles, systems theory, contextual work delivery delivery framework, etc are combined,therefore not a simplistic 1+1 but 1+1 representing differentiated strengths and qualities of strategy, work, people etc., the mathematics of OD and HR then becomes robust and alive. This is research we, Kontextit, have been doing for the past 14 years and now ready to go to market with. Difficult to explain all the nuances in a short message reply. pieter.marais@kontextit.com

Prasad Kurian said...

Thanks Pieter. The tricky part would be to have an empirically valid framework/reference (that is applicable across contexts) against which deviation can be measured. Also, in this approach the underlying definition of quality becomes more like 'absence of deviation' as opposed to 'presence of value'! Sometimes, the value of OD is to enable people to do sense-making or even to 'see the world in a grain of sand' (which is more poetic than logical)!

Mehul @ Search Results Media said...

As a part of HR industry this a very nice article.. Thanks for sharing it sir.

Unknown said...

Very interesting and worth indulging in a little more.
Would like to get in to contact with you.
- Dr. R.Santhanakrishnan, Ph.D.
Coimbatore, tanasan09@gmail.com