Advertisement

The Next Frontier for Human Resources? It’s People Analytics

Article main image
Feb 9, 2015

 A new way of managing talent is beginning to emerge within the human resources world called “people analytics.”

It attempts to remove gut instinct, intuition and human biases from talent management in order to make workforce decisions in an evidence-based and data-driven way.

In other words, we want to ensure that we make decisions around employee selection, development and separation based on criteria that we know matters – because we have correlated them statistically with outcomes – and ignore the criteria that don’t.

Transforming the world of HR

This approach was popularized by the Oakland Athletics when they used data and analytics to help scout and evaluate baseball players. But its roots run much deeper than that, as predictive analytics have effectively transformed entire industries like marketing and finance.

Whereas products were once marketed based on the gut instinct and intuition of marketing executives, data and algorithms are now used to predict consumer behavior with a high degree of accuracy. The result is a more data-driven decision-making process around how, when and where to place ads that increase their yield by an order of magnitude.

It is virtually indisputable that predictive analytics will transform the human resources world and the way that human capital decisions are made. What is somewhat surprising is the fact that we’re relatively early in this transformational process in spite of the fact that these tools and techniques have existed for a number of decades and have already been successful in disrupting a number of other industries.

Decades ago, FICO scores transformed the credit industry and moved the decision of whether someone was creditworthy out of the hands of individuals and into the hands of computer algorithms that had been validated against outcomes. Likewise, advertising used to be done in a fairly untargeted way where random ads would play on your radio, appear on your TV or show in your browser window. Now the Internet has completely transformed how we market to consumers such that you are presented ads for goods that you are much more likely to purchase because of your stated profile preferences, previous habits or websites you’ve visited in the recent past.

The 4 stages of people analytics

This transformation that has completely revolutionized other industries is gradually making its way to the human resources world. And as organizations go through their own journeys, I find that there are typically four (4) stages of maturity associated with people analytics that they exhibit:

  1. Ad hoc question answering — This stage of maturity involves pulling together Excel spreadsheets or other data sources on a fairly ad hoc basis in order to answer targeted questions. The analysis is typically standalone and cannot be easily replicated to answer another question. Some examples: Who are my best and worst managers? Is my referral program working?
  2. Retrospective data analysis — This stage of maturity involves pulling data into a centralized location to do retrospective analysis through basic statistical techniques like correlations, line graphs and others. This process may still be fairly manual in nature or may be accomplished through the use of business intelligence platforms and/or human resource dashboards.
  3. Predictive analytics — This stage of maturity involves pulling data together into a single, centralized location – much like stage two – but then utilizing more advanced econometric techniques (e.g., multivariate regression) in order to engage in predictive modeling and “what if” analysis.
  4. Experimental design — This stage of maturity involves engaging in experiments and A/B testing to elicit true causal drivers within the workforce, studying which interventions are most effective, and then applying them to the non-experimental population.

From the hundreds of conversations that I’ve had with large companies and prospective clients, my personal assessment is that the vast majority of organizations are currently somewhere between stages one and two. They’re analyzing their data in a retrospective manner in order to answer specific questions or produce regular reporting, but not engaging in predictive analysis or running experiments.

The next frontier for organizations

Predictive analytics, or stage three, is really the next frontier where organizations are hoping to land with people analytics. It provides the ability to engage in simulations and “what if” analysis in order to project forward the changes that you plan on making to your workforce.

What if you increase wages across the board by $1 per hour? What if you were to increase paid time off from two weeks to three weeks? Before you make far-reaching changes to your workforce, you can use retrospective data analysis to predict the impact that it’s likely to have on employee tenure and performance.

While analyzing your existing data is a low-risk approach to adjusting the knobs and levers affecting your workforce within a black box, it also has the drawback that it isn’t truly causal. For example, if you find that wages are associated with longer tenure, it’s impossible to determine whether people performed better because they were paid more or if they were paid more because they performed better. That’s why some forward-looking companies like Google are running experiments to establish truly causal relationships.

Recently, Google ran an experiment called Project M&M on the candy that they famously make available to their employees. They put the candy in opaque containers and the healthy snacks like dried figs and pistachios in glass jars to see how consumption habits would change. In the New York office, a staff of 2,000 consumed 3.1 million fewer calories from M&Ms alone over a seven week period.

This is just one example of the sort of data analysis and experimentation that market leaders are beginning to explore with respect to their people.

A data-driven, evidence-based approach to HR

Of course, we obviously need to recognize that these are people and so we can’t easily tinker with things like their compensation or benefits. But if we’re deciding on a new referral policy, a new layout for the office, or a new schedule for Happy Hours and company events, doesn’t it make sense to use this as an opportunity to design an experiment, roll it out in a staggered fashion, collect the data and determine what worked the best? This is where I think the industry is heading in much the same way that the A/B test is the gold standard for marketers everywhere.

It’s an exciting time for the human resources world. People analytics is slowly transforming the way that we hire, develop, and promote or separate our workforce in a data-driven and evidence-based way.

The organizations paving the way tend to be early adopters, innovators and disrupters like Google. I think the next five years are going to see a sea swell of transformation as others jump on board and realize that they absolutely must adopt this same philosophy if they wish to compete within a service-based economy where human capital is the most valuable strategic asset.

Much like in Major League Baseball, where there is hardly a team that would be caught dead ignoring the sabermetricians and using the gut instinct and intuition of their scouts to evaluate talent, we’ll soon see a world in which companies either leverage people analytics or fall further and further behind their competitors.