THANK YOU FOR SUBSCRIBING
Data Science and Human Resources are still a new pairing to a lot of companies who might already have well-established data science capabilities in other departments such as Marketing and Research. As the existence of advanced analytics in HR becomes more prevalent it is important to discuss how to best work with such a team and what they can offer. Famed psychologist and pyramid-based graphics enthusiast Abraham Maslow is credited with saying “I suppose it is tempting, if the only tool you have is a hammer, to treat everything as if it were a nail, "and this sentiment is not misplaced when talking about data science.
It can be tempting for leaders to want to lean entirely on data and embrace the information revolution to inform their decisions. However, this is often unrealistic and detrimental to both parties. Firstly, not all data is created equally in terms of quality and depth, and secondly not every question can be answered solely with data – hardly any question can be answered this way in fact. This misunderstanding can lead the Analytics team to suffer reputational loss when they cannot produce something in-depth for decision making, and the business leaders are left with suboptimal recommendations and solutions.
The secret sauce that brings all the disparate pieces together is communication
The best way to use an HR Analytics team is as part of a solution or strategy, not the only component of it. At Shell we say that data is only one part of the equation. Experience, intuition, context, and classic information gathering tactics like surveys and assessments are as important for HR projects as more traditional datapoints. The secret sauce that brings all these disparate pieces together is communication. The analytics team and the business stakeholders must speak frequently and in depth to understand both what the problem really is, how the data science team can contribute, and what the solution or end-result will be. A dialogue needs to exist for both parties to make full use of their expertise - data and business knowledge. Lastly, both teams should always strive for simplest solution possible. Neural networks and NLP are powerful tools, but they are often more complex and less interpretable than desired. A simple regression is just as good as an AI solution if they both help the business advance towards its goals.
To bring it back to Maslow, it can be tempting once you discover the existence of an HR Data Science team to ask them to solve every problem with a complex data science solution. However, it is more important to have strong communication between analysts and the business to establish a deep understanding of the problem, make use of outside research, previous projects, and historical in-house knowledge, and then finally when reaching into the data science tool bag to make sure that the simplest and most appropriate tool is being used each time.