Enhance your HR with People Analytics

"The world is full of obvious things which nobody by any chance ever observes"

Quoted Arthur Conan Doyle, which was well fashioned by his globally celebrated character- Sherlock Holmes through the art of deduction. If this 'high-functioning sociopath' taught us anything, then it is that people inadvertently leave traces of their personality out for observation. If these cues are interpreted correctly with ample data, one can explain an individual's behaviour, and even make behavioural predictions in some situations.

So, how do you observe these traces of people personality?

"Well, it's elementary, Watson!"

Observing people data to make inferences that either explain your workforce behaviour, people performance, learning trends or predict areas of unrest, flight risk, talent pools is basically People Analytics, which took us about 130 years to adapt into corporates when Mr Holmes had already yelled 'elementary' about a million times a century ago.

Following are the types of HR Analytics

For ease of understanding, we will classify the whole spectrum into 2 simple classes-

1. Explanation: To explain trends/ decisions using the data that is already present or making sense of things. Think of it like understanding why the manufacturing batch failed at a large scale and finding out what variables are the most impactful to drive that failure. Descriptive and Diagnostics Analysis would fall into this category.
2. Forecast: To predict what could be the trend/ outcome given the data from many past results from similar situations. This is similar to predicting stock market trends, sales forecasting or even weather forecasting. Predictive and Prescriptive Analysis would fall into this category.

Many business functions have been using analytics for decades now to make progress, and in the flow of these numbers and equations industries are now adapting analytics in Human Resources as well, especially to aid workforce planning and decision making through logical and numerical arguments.

The adoption of an inordinately mathematical field in an area which was historically commanded by perception, assumptions and some measurement is challenging but most definitely not impossible. It is important that the leadership in any organisation, especially the CHRO is visionary and willing to experiment and accept this evolving field.

Having noted that, the acceptance of People Analytics (PA) has not been discouraging at all as per this Deloitte report of 2017. Per this report, more than 70% companies say that they consider PA to be a priority.

So, where do we start?

Fortunately, we don't have to invent anything new. Much inspiration can be drawn through Project Oxygen by Google that determined the mentorship methods of their best managers and later used these methods in coaching sessions to improve the output of low performing teams. (Here is a Harvard Business Review Article on Google's Project Oxygen)

And if you are from the Netflix generation who prefers video content over readin, here is a video for you:

Now that you know how the tech giant Google did it, you have a potential goal in front of you. By no means is this goal achievable overnight, especially not without clean data or a great team.

Wouldn't it be easier if someone had a step by step approach to getting started on this? Well, you are in luck, Watson, because someone already did- McKinsey!

This Stairway to Impact Model is condensed perfectly by McKinsey's article on 'How to be great at People Analytics'.

Now that you know these steps, it is imperative to address the elephant in the room.

Just because PA is the global trend, and is potentially "cool", many companies want to try their hands on it.

When asked the simple question of 'What business problem would you like People Analytics to solve for you?' the answer usually is 'What can you do with People Analytics?'

This is a big problem, because there is no business challenge to solve. The most important and irreplaceable step in any form of analytics is knowing what problem you are solving. So, before we even go to McKinsey's step 1, we need to discuss step 0 thoroughly- What is it that the company wants to do with People Analytics?

So, my dear Watson, if you are ready, here is a non-comprehensive list of Projects you could do with People Analytics.

1. Explanatory Projects

• Explaining performance rating of your employees using Multivariate Regression Model
• Explaining training needs per demographic using historical data of training
• Finding most useful and most profitable trainings using ROI and Kirkpatrick Model
• Explaining employee sentiments/engagement through- (i) Surveys (check for Cronbach Alpha for validatiy) or (ii) Network Analysis (most important and most advanced)
• Explaining depth of network to check the speed of decision making through interaction matrix. A top down approach is the best foot forward for this
• Explaining the effectiveness of communication by identification company influencers
• Finding out the correlation of various variables with performance to determine the key indicators of productivity. Use pseudo R^2 and p-value to estimate if your regression results and conclusions are acceptable
• Explaining employee burnout behaviour through time of response and engagement rates
• Explaining promotions using Binomial Regression
• Integrating behavioural capital with finance and HR data using Game Theory

2. Forecast Projects (more advanced)

• Flight Risk to predict which employee might quit
• Low productivity impact on BU performance. After finding out the variables that dictate productivity, predict that given the set of those conditions, what is the likely outcome of the BU performance
• Using outlier data as influence, predict the behaviour of the organisation in cohorts in adoption of a new policy
• Predict future skills based on job descriptions as a form of natural career progression
• Predict hotspots for talent attraction
• Predict the best trainings needed to upskill your staff that will also have a high ROI

What fields of HR can use People Analytics?

However, in order to do any of the above, the first thing that you need is the problem statement; the second thing you need is clean data and audited data that is up-to-date and has as few mistakes as possible, because a part of this data will be used to train the model and the other part will be used to predict with certainty; the third thing you need is a great team- a leader who understands mathematics, HR, psychology alongside a team of data scientists and HR professionals who can collaborate with business leaders, IT, finance and HRBPs to do their magic.

Here is an example of self-audit-

This is not to be confused with HR Operations or HR Technology which deals more with data storage, HRIS and processes.

So, now that you are an expert at elementary, Watson, would you like to start?

For decades now, HR has been gravitating towards psychometrics, modern history, organisation psychology and philosophy, but let us now introduce mathematics and psychophysics to this domain.

In 1943, Kurt Lewin proposed an equation to the world of psychology, now known as Lewin's Equation, combining behavioural identification and prediction using mathematics. His equation states that behaviour is a function of the person and his environment.

As Lewin started with just B=f(P,E) , we can begin small too, by acquiring the right data, making explanatory models and then moving to automation, followed by prediction and forecasting.

In a world where your employees are no longer interested in filling taxing surveys, or have attained the skill of answering in a way that pleases the employer, getting correct results by asking questions directly to the employees could potentially be a waste of time, energy and resources in a few years. The future of sentiment analysis lies in non-invasive methods of measurements by mere observation of behavioural cues. There are already proven techniques to do this and academic articles explaining the methods, the only thing now stopping us, is the vision and acceptance that the world is moving and it is moving fast!

The pandemic has definitely changed employment giving it a massive transformation. Leaders who believed that remote work wasn't possible were forced to experiment with remote work with now proven results that people can and will work without babysitting.

Here is a great article Worker-Employer Relation Disrupted from Deloitte explaining the pandemic changing the perception of employees and HR.

On closing notes, I don't have anything better to add than what Google People Analytics stickers would say at one point in time-

" We have charts and graphs to back us up, so, f*** off "

Employee engagement, payroll, appraisals and other processes are not just a tick in the box, a happy employee, is a loyal employee and is more productive. Use historic data to your advantage to serve your employees better.

-Ajinkya Bhasme