Skip to content
A Magic Tuple of Prescriptive Analytics in Workforce Development
In my previous article “AI and predictive analytics in workforce development (WD) are trends in the future”
the market tuple was introduced as a data structure
“industry — location — skill — proficiency level — time”
in the 5-dimensional system of statistical research.
However, today I want to correct this tuple. I realized that “management level” is a very important axis for analytics in HR and links all factors with a financial indicator.
Currently, I see the market tuple as a system of coordinates for predictive analytics in HR as a 7-dimensional hypercube of:
In this article I would like to explain the meaning of market tuple and introduce some metrics which are useful for people (workforce) analytics.
First, what is a tuple? According to the Oxford dictionary a
in computing is “a data structure consisting of multiple parts”. Second, you can understand the market tuple as a point into an n-dimensional system of coordinates where coordinate axes are the above mentioned factors.
defines industry term as a distinct group of productive or profit-making enterprises. In order to building representative statistic in people analytics we need to know an industry context. For example, the CEO in the pharmaceutical sphere has not the same skillset as the CEO in a software development company. Of course, some of their skills are equal. It means, that
classification of skills has to contain information about skills in relation to industries.
International Standard Industrial Classification of All Economic Activities
seems like the most well-grounded solution.
Clearly, workforce development depends on location. For example, workforce in China and workforce in France are absolutely different with specific development traits. A more clear example — astronauts in Space vs astronauts in training centers. Unfortunately, I couldn’t find a classification which accounts for workforce development on Earth and out of Earth. This classification could be developed starting from the root concept — the Space. The next level — galaxies. After galaxies we have to define the planetary systems (Solar system). In this classification the Earth will be a child (subclass) of the Solar system. The Earth as an umbrella term has to be represented by continents, countries and so on.
If the “industry” axis helps to define a sphere of economics, the “location” helps to account the context of place.
In the above mentioned 7-dimensional system a skill is a central factor for
in workforce development. I can assume that head of a company as an occupation existed in the past, exists today and will exist in the future. But a skillset for the same job position depends on time — some skills become more actual with time, some skills — lose their actuality. In this article I do not separate skills on soft (competences) and hard skills.
As I wrote in an early article “Semantic interoperability as a basis of Meaningful Analytics in HR”
, the European Commision is in the process of creating and validating European Skills, Competences, Qualifications and Occupations classification. It is a very appropriate and useful initiative that has to be supported by analytical experts. As I pointed above,
some skills are related to an industry context but it is only true for hard skills because competencies are cross-industry skills.
We often read following titles — senior developer, middle/regular developer or junior developer in vacancy descriptions of software development companies. What does it mean? It means what a company defines as an expected level of qualification. I will not reinvent a wheel and see that the Competencies Proficiency Scale
of the Office of Human Resources at the National Institute of Health can be successfully used for HR analytics as well. It contains the scores, names and descriptions for each proficiency level:
— Not Applicable;
— Fundamental Awareness (basic knowledge);
— Novice (limited experience);
— Intermediate (practical application);
— Advanced (applied theory);
— Expert (recognized authority).
In this part of my clarification I have to say something about years of experience. I do not include it as an axis in my system of HR coordinate. Why? Because it is rather a passive restriction for obtaining some position.
More interesting and important is the awareness on the progress/speed on reaching a higher proficiency level. What is a expected time for any individual to achieve the next level of competence in a particular skill? This metric permits to identify individuals that have a high potential for professional development.
F. John Reh in his article “The Use and Creation of Position Grade Levels”
writes: “The creation and maintenance of a system of standardized employee grade levels helps ensure fair compensation for the same level of work across different departments and divisions.”
He introduces and describes the following levels
Level A: Entry Level Individual Contributor;
Level B: Experienced Individual Contributors;
Level C: Managers and Senior Technical Professionals and Individual Contributors;
Level D: Directors;
Level E: Vice President/General Managers.
And we also should know on the progress of a person in achieving the next level in career development.
If “Proficiency level” often indicates high-professional candidates, “mangement level” often indicates high-potential candidates.
Time and Cost
Time and cost are indicators of a market. Time is important because it provides you with information of a particular required skill set, knowledge at a certain point of time. Since time is constantly changing so are skill sets as well as supply and demand. Supply defines a wished compensation by a candidate for a particular skill depending on industry, location, proficiency and management levels. Companies define through demand a price for a particular skill depending on the same factors.
If we can represent an individual as sum of his/her skills and know the demand and supply for each skill, we are able to compute a labour market value for this person.
Some models and metrics
Earlier, I wrote about the speed of progress in proficiency level and management level. Let’s remember that the proficiency level scale has 6 items and the management level scale has 5 items. It is obvious that we will not indicate “not applicable” as a proficiency level on the management level status therefore, only the next 5 levels apply in this case.
Development of workforce is our issue that defines our focus on the model of development. So, we have 5 levels for proficiency and management and it means that we can define a dataset of pairs “time to achieve -> level”. For any market we can use a polynomial interpolation with degree of the polynomial that equals 4 because “If a set of data contains
known points, then there exists exactly one polynomial of degree
-1 or smaller that passes through all of those points.”
Mathematicians know that the “polynomial interpolation, often (but not always) provides more accurate results than linear interpolation.”
The polynomial interpolation is also called a polynomial model.
If we have a polynomial model for market and polynomial model for particular individual we can answer on the question about future development of a person. Because we can find similar cases from others and perform predictions based on precedent. This approach is often used in medicine — doctors can predict on disease progress on the basis of similar cases.
Of course, if we have knowledge about supply and demand on the labor market we can use applied statistic for prediction of employment termination and calculate a confidence for this.
Described approach helps to build a planning process for education in a very pragmatic approach. Because it lets to know:
which candidate’s skills are weak and which are strong?
which skills have to be improved?
what is the expected time to obtain or improve some skills?
how much will the necessary education/training/course cost?
what is the risk of terminating employment after education
what is the expected time of burnout and a lot of other metrics.
Erik van Vulpen
in “The basic principles of people analytics. Learn how to use HR data to drive better outcomes for your business and employees” writes (
, page 27): “On average,
organizations spend 1200 dollars per employee on training and development
. This amounts to a yearly spending of 70 billion dollars in the U.S. alone.
That’s over two times the amount of money needed to end world hunger. Do we actually know the impact of these investments?”
The answer is NO, today we do not know! However, Naomi already has a scientifically proven method which can answer this question!