Blog

by Jason Ezratty and Michael Benyamin

Although the volume and quality of data generated by HR systems is escalating, data science in HR is still relatively immature. Challenges abound. For one, there’s a lack of clarity around what “good data” looks like, making it difficult to understand how to deliver people analytics services and products to customers.

Other common challenges include:

  • Limited in-house expertise.Lack of all the necessary skill sets required, ranging from business acumen to data management to data visualization, storytelling and programming to train and deploy machine learning algorithms
  • Resource constraints.Inability to take on additional time intensive projects because there aren’t enough resources to meet capacity
  • Lack of tools.Without automation technologies, manual data refreshes could take weeks
  • Unsure of data quality.Analysis findings are questionable because there is no process in place to continually improve data quality
  • Difficultly with C-level buy-in.Without experience using data to make decisions, not all clients understand the value, making them unreceptive to new delivery models

Regardless of the challenges, interest in HR Data Science is on the rise. Over the last five years, HR has become increasingly sophisticated in how to leverage data beyond the basics of who, when, how many and how much. New state-of-the-art technologies enable HR systems to go beyond what programming logic can be committed to code and glean insights learned from the data, such as predicting candidate fit and employee retention risk.

So how can every organization realize the benefit of HR Data Science given variables like available staff and level of in-house expertise?

Understanding the Business Focus is Key

Beyond conventional headcount and turnover reporting, not every organization’s demands for people analytics are the same. A company going through significant M&A activity, for example, is going to be focused on analyzing how to combine strengths, retain key people and eliminate redundant work. An organization in a stable environment may be more interested in broad performance management, leadership pipeline and employee retention.

In addition, there can be a disconnect between the perceived investment in people analytics and its value. Questions include “how is ‘this’ different than what we already have” and “how do we know that ‘it’ will create the value we need”. As Albert Einstein said, “We cannot solve our problems with the same thinking we used when we created them.”

 

Ultimately, companies need to change thought processes and adopt the tools necessary to deliver HR services efficiently and effectively. HR needs to explore what work can be automated, where it can benefit from more intelligent decision support and how HR can best use the technology as business demands grow over time.

Various machine learning and data analysis technologies, for example, enable HR teams to adopt AI-driven chatbot technology in talent acquisition – either for improving the hiring manager or candidate experience. Products like Beeline Virtual Assistant allow hiring managers to describe the role they are trying to fill and the virtual assistant performs all the transactional work. Some organizations are using virtual assistants to help gather initial candidate data, schedule interviews and maintain an ongoing connection with candidates. And, then there are specialized apps, like Talent Data Exchange(TDX) from Brightfield, that offer AI-based decision support for labor cost and relationship to quality for contingent labor.

How to Determine the Best HR Services Delivery Model for Every Organization

Most people analytics teams start small – often just one to three people. These teams may excel in some of the required capabilities for people analytics but rarely all. Our view is summarized into top three capabilities:

1) Understand the problem

2) Apply science to make data-based decisions

3) Use change management to help the organization understand the data and incorporate it into routine decision making.

Organizations without strong change management capabilities may find additional hurdles. People analytics teams may face roadblocks if they can’t illustrate the value of their findings to executives, are unable to understand the full context of the problem or lack the relevant experience to know which practical solutions could impact data insights. Gaining awareness of your organization’s maturity and readiness for analytics, how to respond to current and future demands and having a business leader’s mindset of the analytics function are critical for delivering anticipated value in people analytics.

Learn more by reading an accompanying blog post, The Process of Determining the Right HR Data Science Delivery Model.