HR Analytics in the driving seat

Given that many companies are starting to put words into action and People Analytics is gaining momentum, I thought it would be a good moment to write how leaders should guide their HR Analytics team. This article is addressed to a fictitious Head of HR Analytics of a big multinational company.

Step 1. Explore your own data science profile and that of your team

Data scientists are a valuable asset to any organisation, but unfortunately all-round data scientists do not exist. Therefore you need to understand your own skills profile and that of your team. Online tools exists to test skills profiles. It will help you to objectively identify your own profile and that of your team, establish some training paths, monitor the learning progress and the skill sets that need to be found in potential new recruits.

Thinking that HR, legal experts, business representatives and IT, constitute already a multidisciplinary team is reductive. Data scientists are multi-dimensional, so different data scientists are needed to constitute an all-round HR Analytics team.

Step 2. Clarify why HR Analytics is needed

Like in all change management exercises, it is necessary to be clear why you are on this journey. Clarity improves chances of success and resilience with the inevitable problems along the journey.

Data Science is required to maintain competitiveness in an increasingly data-rich environment. 90% of the world data were created in the last two years alone. In 2020 it is expected that the same amount of data will be created in a much smaller fraction of time.

As per a document recently published by Booz Allen Hamilton (second edition of The Field Guide to Data Science), a 5-6% performance improvement can be observed in organisations making data-driven decisions.

The tangible benefits of data include better opportunity costs and enhanced processes: Opportunity costs rise when a competitor implements and generates value from data before your own company. As a result of an increasingly interconnected world, huge amounts of data are being generated. Data Science can be used to transform data into insights that help improve existing processes.

Step 3. Assess your own organisation’s maturity model.

I always recommend collecting information on the current status of the HR Analytics maturity model. The results will help in setting priorities. The siloed-HR-data approach might become untenable in the wake of easiness of acquiring information and the sheer amount of HR data available. Smart organisations tend to consolidate their data and gain business insights from them.

Step 4. Take the geek approach in all your tasks

Every task or project, you are given, try to fix it for good, as per the famous geek versus non-geek approach.

HR Analytics is an investment, so keep investing relentlessly in it. With your initiatives, you are going to create an ongoing and substantial long-term demand for HR Analytics in your company.

Step 5. Aim for long-term credibility

HR Analytics has a high hygiene factor. When you are good at data cleaning, nobody really seems to care. But if you are not, you will be in deep trouble, contributing substantially to the loss of credibility of your HR Analytics function. Data cleaning and exploratory analysis will occupy the vast majority of your time with little time left for model discovery. Setting up a machine learning model is a good start, but increasing its accuracy and usefulness requires an increasing effort.

Step 6. Spend a lot of time to train yourself and ensure the same also for your team

Personally, I spend a lot of time training myself to learn R and statistics. R is a leading tool for data science. One day I will also study Python. Tools like Excel are great for doing quick data explorations, but the work gets difficult to reproduce and therefore end results cannot be double-checked.

Step 7. Get exposure, compete in competitions and liaise with peers and experts in the HR Analytics arena

We live in a dynamic world. One has to keep constantly up-to-date with the technological evolution. You have to talk with experienced professionals, start participating in discussions, meet-ups, hackathons and join groups. To be truly successful in HR Analytics, you have to test yourself against your peers from time to time, for example by participating in a Kaggle competition. I would keep an eye especially on competitions relating to text mining and graphs presentations, as I see those techniques to be very valuable in HR.

Last year I spent two months of my free time analysing a practical business case, just to practice with data analysis and gain experience with R. I checked if it was possible to predict pregnancies on the basis of woman’s age and any previous children, only to conclude that it was not possible to predict such events in a company setting with so little information. In the end, I had to fall back to a simple regression model. Nevertheless, it was an excellent experience, which allowed me to get exposure on how to build a hypotheses model, verify it with the analysis of variance, discriminant analysis and decision trees. In predictive data analysis, the search for the ‘best’ predictive model can be rather elusive sometimes.

Step 8. Disrupt or be disrupted

At the end of the day be ready to take some radical decisions or recommend those to your company. “The riskiest thing we can do is just maintain the status quo”, said Bob Iger, just to mention one of the many inspirational quotes inviting to take action. Your insights will start making an impact on the business.

Here I would like to mention just two radical decisions taken by big companies in HR recently: GE, Deloitte and Accenture changed their annual performance review in favour of continuous, real-time programs. Ernst & Young, one of the UK’s biggest graduate recruiters, announced it will be removing the degree classification from its entry criteria, saying there is “no evidence” success at university correlates with achievement in later life.

More of these radical decisions will come in 2017.

Remember that “52% of the Fortune 500 companies have disappeared since 2000″ and that “the digital disruption is more than just a technology shift. It’s about transforming business models and how we engage” to quote R Way Wang.