HR Analytics: Taking HR data beyond the headcount

By Stella Voules & Jo Billing

10minute read…


 

In a previous post, we discussed the extensive impact that several up-and-coming technological breakthroughs are having on HR as we know it.

A prominent theme was that Big Data and the Internet of Things (IoT) will pave the way for a golden age of powerful analytics, giving HR functions more insight than ever before; with the added benefit of reallocating time to more important tasks as a result.

IBM reports that every day 2.5 quintillion bytes of data are created; 90% of the data available today was created in the last two years.

Suffice to say, we have more access to data than ever before. This post will discuss what that means for us.

 

Moving beyond simple Descriptive Analytics

What’s our headcount? andHow does it compare against budget?these are two inevitable questions in any organisation, and they will have you spending most of your day (or week) downloading, cleaning and validating data. It takes a lot of effort to establish real, actionable insights – but what’s done with the information you have worked so hard to provide?

Typically, HR teams do have access to some business intelligence tools that provide descriptive data for reports and dashboards, as business intelligence software commonly provides a descriptive analytics platform. While many would agree that the current iterations of HR software are far from perfect (and a little expensive) – it is worth the investment as you’ll avoid drowning in Excel spreadsheets and simply have more time to have datainformed conversations with leaders about what’s going on.

The analytics landscape is a broad one. Most often, you’ll come across these approaches: descriptive analytics, predictive analytics and prescriptive analytics.

At the level beyond descriptive data, the analytic capability of HR Business Partners and many managers to understand reports, make insights and manipulate data is limited.  Creating customised analytics needs some savvy coding and statistical analytics knowledge to get the algorithm right.  Rather than building large data analytics teams – one solution is to invest in Workforce Intelligence software which packages the coding and analysis up.

Predictive and prescriptive analytics tend to operate in the way a doctor’s diagnosis does – the doctor tells you what’s wrong, and if you take some medicine you should be cured – that’s the prediction.  To cure the symptoms, you are then provided with instructions of which medicine to take and when that’s the prescriptive component.

Prescriptive analysis has always been a rarity in HR, as human behaviour is of course more complex and less predictable than tangible assets. In the future, this is likely be different. This is where neural networks, deep learning and Artificial Intelligence are starting to show some promise in what might be possible with HR data.

Predictive Analytics: The Holy Grail of HR


Good analytics are the meeting of art and science.

To use the analogy of a journey, statistics is the engine and vehicle that is used to travel, while analytics is the plan behind it.  Without a travel plan, the vehicle and engine can’t really take us anywhere; we’ll simply drive around aimlessly.  

For example, descriptive statistics of exiting employees will sometimes highlight a turnover issue – however predictive analysis will often show that managers overestimate the talent in their team and underestimate the flight risk of people leaving.  Without asking the right questions this turnover issue would never properly be addressed and the conversations about what’s going on would go in circles. That makes predictive analysis a powerful capability.

Predictive analysis can help HR become more strategic and stop putting out constant spot fires.

 

Prescriptive Analytics – More than a crystal ball

The real prescriptive power is, however, still largely in the world of the numerati – its in the emerging Expert Systems / Neural Networks / Deep Learning analytics. These are forms of machine learning where the analytics mimic the thought process in a brain – by moving data across points they call synapses and neurons. Modeled loosely on the human brain, a neural net consists of thousands or even millions of simple processing nodes that are densely interconnected, they analyse the hidden layers of connecting decisions between point A & B.  These analytics are not necessarily new, however what is new is the size and scope of the data that is available to mine due to recent technology.  It is the access to this deep and abundant HR data that holds promise for the future of HR.

These advanced analytics can connect your HR data with revenue, profit, and customer satisfaction. They can predict (with some accuracy) who will leave–however, there are ethics on acting on this (see our article on Digital Ethics).

These analytics will not replace human intervention: they won’t tell you the one clear course of action to take, particularly when dealing with data that has feelings – people data.

Some advice from OrcaEyes CEO, Dan Hilbert, “Statistically, predictive loss analytics can be no more than 25% accurate. It’s simple math and sampling size. Whereas SonarVision shows the 4 primary reasons for potential employee turnover, use of Predictive individual employee Loss Analytics, needs to be used with extreme caution in the real world. This intelligence should be used as an “indicator,” not a fact. And the use of this data to proactively intervene must be performed with high disparate impact rigor. To reduce legal liabilities, these interventions must be done with a program designed to ensure all groups are treated fairly.

 

Some final thoughts

High level analytics are far more valuable than naming individuals, as they allow HR to develop thoughtful, refined, long-term programs to reduce organisational issues such as:

turnover by targeting root causes
systemic issues driving misalignment of applicant selection and talent management
connecting employee and customer engagement and understanding the drivers
risk assessment.

These advanced analytics will not only allow you to better predict the future (which assists decision making) but also help you learn from the past. Managers should then be able to self-identify the risks.

Prescriptive analysis is obviously incredibly powerful, and while it is still slowly filtering through organisations, it is worth the early investment. Creating game plans within HR departments to start implementing current versions of software capable of performing this, or to lay the foundations for a strong presence, will likely be an investment worth making for driving productivity and competitive advantage.

In the future, we will be able to develop a deep understanding of the connections and drivers in relation to our people.

It is at this frontier where human behaviour starts to become a little more predictable and HR departments will be able to genuinely and deeply influence business decisions.

 

Here at JOST & Co. we are on the pulse of breakthrough technologies and strategies for HR departments to keep up with this fast-moving workplace and technological change.

If you have any ideas, questions or queries, feel more than free to contact us. We’re always up for a coffee!