HR's Dying Breed of Data Gurus

Prepare for the future of HR analytics -- and the decline of data scientists -- by moving carefully through the changes and answering some difficult questions.

By Wes Wu, Clementine Wong and Anthony Scaglione

In a highly competitive talent market and dynamic business environment, HR professionals are now better equipped to gain insight into workforce behavior and trends, thanks to advances in data-analytics capabilities. Companies can more easily redefine their HR operations and capitalize on new efficiencies and strategies, but getting from theory to reality -- in a marketplace increasingly saturated with new solutions, and with technology rapidly evolving -- can be a struggle.

Consider the role of data scientists, who have begun to emerge as the heart of effective analytics systems. At a basic level, these are the people who corral and manage your data to derive meaning from it and answer strategic questions. They accomplish this through understanding and merging data sources, making them consistent, structuring mathematical models, and presenting visualizations and the resulting insights, among other tasks. For many organizations, building and maintaining an effective future-state analytics model in your HR department requires hiring data scientists who know the specifics of the business and the underlying data.

But finding data scientists with HR backgrounds is a challenge -- and even then, a number of factors may be reducing the value they can provide. By asking merely operational questions, many HR leaders aren't fully exploiting the capabilities enabled by analytics, focusing on past outcomes instead of future possibilities and strategic goals. Meanwhile, data scientists' traditional areas of focus have become more easily and cost-effectively addressed by increasingly sophisticated and commercially available HR-technology offerings -- a trend that's expected to accelerate over the next three to five years, thanks to advances in artificial intelligence, machine learning and software robotics.

This article examines how necessary data scientists are in the analytics landscape and the trends that are revolutionizing the field, considered within the lens of how HR organizations can deliver optimal insights and results while preserving their long-term vision and cost efficiencies.

Talent and Solutions Landscape

Analytics come in three varieties: descriptive, predictive and prescriptive. All too often, HR organizations -- and their data scientists -- rely on descriptive plans, which look only at what happened and why. Predictive uses data to look toward what could happen; and prescriptive, the most advanced on the spectrum, takes everything a step further, mapping scenarios for what may happen, when and why, corresponding actions to take, and the results from potential decisions and opportunities.

Successful future-state analytics plans are built to look forward, and that requires asking the right questions and relying on the right people and tools. For example, asking, "How long has it taken to hire someone at a certain rank?" -- a descriptive question -- likely requires only a basic solution, with no data scientist. But you gain more strategic value from asking, "What is the success profile of the best candidates for my organization?" and "What makes new hires successful, and what convinces them to stay?"

From the beginning, companies have focused their analytical efforts where they have the most data that could have the greatest impact: about their consumers, their finances and their supply chains. As such, these areas have seen more interest among students who are pursuing data-science degrees and jobs. But fewer are focusing on people data, covering recruiting, talent management, engagement, retention and the other primary areas of focus that make up human resources. Exacerbating the situation, in those organizations where HR data scientists are employed, HR professionals and business leaders are generally not asking the most meaningful questions. Instead of seeking insight and foresight, they are largely asking for hindsight data. The result: Data science has not focused on HR-oriented problems and HR-analytics teams often find themselves working on topics that prevent them from delivering on their full potential. 

These days, many of the new HR-analytics vendors have specialists on their staffs to work with HR professionals and business leaders, and they offer product suites with many prebuilt reports and visualization capabilities. In some cases, vendors' Software-as-a-Service tools can benchmark your company against their other clients. Many of the newest vendors have deployed their own data scientists to approach the right questions, allowing HR to gain insights without hiring their own data-science staff.

So, even if you do hire your own data specialists, will you be able to find a data scientist with HR-domain expertise, and how long will it take him or her to gain a sufficient understanding of your organization to be useful? And will he or she end up being underutilized on more descriptive and backward-facing efforts? For some companies, the base amount to invest in in-house talent and tools is too high to be cost-effective -- and they could become rapidly obsolete on top of it.

A Changing Reality

In March 2016, two opponents squared off in a game called Go, a Chinese game of strategy first developed thousands of years ago known for being vastly more complex than chess. On one side was a man who had won 18 international titles. On the other was a computer program. Each side had one minute to make its move. Millions of fans were watching -- and Silicon Valley, hungry for the next big innovation, was paying close attention too. In the end, the best-of-five series marked another decisive victory for artificial intelligence.

The computer program relied on a vast catalog of competitions between the game's top players across history and, through machine learning, sifted through that trove of data to uncover winning patterns. In other words, it wasn't explicitly told how to win, but instead was set up to use its advanced processing powers to learn how best to win on its own.

Some of the world's leading tech companies are banking on AI, machine learning and software robotics to fuel the next boom, aiming to produce in code the same kind of learning and inductive reasoning that data scientists are expected to bring to the table. If programs and systems can figure out the context and subtext, relying on their own learning instead of purely on formulaic algorithms, then the benefit of employing a data scientist to do the same is greatly diminished -- especially when the company is focused purely on descriptive, hindsight-focused analytics.

These advances are likely to accelerate. Will you need only half the data scientists two years from now because AI has displaced them? Or what metrics will you use to determine whether to make the leap to a more automated system, and when? You may not know the answers, but the questions should help frame the discussion about what is needed from an HR-data-science team and what's needed in terms of technology.

Go Forward with Confidence

So how can HR organizations create a plan empowered by analytics for the future amid the disruption of the present, in which data scientists are likely to be displaced by evolving technologies? Ask yourself:

Are you looking backward or forward? It's important to realize that just because you have a capability doesn't mean you have insights; the questions you feed into it become the foundation for how useful the answers are and how they can shape an optimal strategy. To capitalize on the full promise of predictive and prescriptive analytics, HR leaders must enhance and upgrade the operational types of questions they're posing to their HR data scientists. Shifting from a reactive mind-set to a proactive, strategic one (as in, from descriptive to predictive/prescriptive) can sharpen your strategy and inform which direction to take.

Are you using your people and tools optimally? As "out of the box" analytics tools from leading HR SaaS applications are continuously evolving, keep abreast of these vendor-delivered offerings, and take advantage of those capabilities to deploy data scientists against higher-order inquiries. And leverage or borrow the advanced analytics experience and insight of other business groups, if available (for example, in finance, supply chain and customer resource management).

Are you up-to-date on how the data science landscape is evolving? Monitor advances in AI, machine learning and software robotics to understand the capabilities of technology as well as the resulting role of the data scientist in your organization. The fundamental question will be whether data scientists are still needed and what role they will play.

Do you need to broaden your knowledge base? HR organizations, including the business partners and vice presidents, need high-level knowledge of what analytics can do. Without it, they'll struggle to ask the right questions and not understand what solution is the best fit.

Do the leaders at the top of my organization know what they want? If the needs are overly complex or if they cannot be articulated, hiring someone in-house or choosing a vendor can be a wasted effort, regardless.

The path forward will likely be challenging. But doing nothing is not an option --  not in today's landscape where competition is unrelenting and technology change is constant. Whether a nimble startup or a more established rival, organizations of all stripes are looking at ways to invest in and capitalize on analytics. Even amid uncertainty, HR can become a genuinely strategic partner by harnessing the power of advanced analytics. The time to start your debate, or to advance it, is now.

Wes Wu, a principal, and Anthony Scaglione, an executive director, work in the People Advisory Services practice of Ernst & Young (U.S.). Clementine Wong is a PAS senior manager for Ernst & Young (Australia). The views expressed are those of the authors and do not necessarily reflect the views of the global EY organization or any of its member firms. Read more at ey.com/PAS.

Sep 22, 2016
Copyright 2017© LRP Publications