Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vivamus convallis sem tellus, vitae egestas felis vestibule ut.

Error message details.

Reuse Permissions

Request permission to republish or redistribute SHRM content and materials.

The Benefits of People Analytics in a Time of Uncertainty

Meta Vice President of People Analytics and Workforce Strategy Alexis Fink shares her perspectives on the state of people analytics tools and the ways organizations can put them to work.

A black and white photo of a skyscraper.

People analytics emerged out of human resources at the turn of the century, allowing us to utilize qualitative and quantitative insights to inform decisions that make organizations more effective and efficient. These insights also provide the information we need to create more fulfilling workplaces where our people can thrive.  

In a recent webinar offered by the SHRM Executive Network, Alexis Fink, vice president of people analytics and workforce strategy for Meta, provided attendees with advice and insights into the world of people analytics. "There is a wonderful opportunity for creativity and really high-quality practice as we try to make our organizations better for our shareholders, employees and other stakeholders," Fink said.

The Challenge of Finding Meaning in Data

In conversations about people analytics, it is easy to forget that data doesn't stand alone; context and analysis are critical, Fink noted. A single data point can have several different meanings, depending on context. This need for context becomes even more critical when dealing with diversity, equity and inclusion data, which can be misinterpreted without broader background information.

You can't use data engineers to extract meaning from qualitative pieces. When we're talking about extracting meaning, you need the expertise of social psychologists and anthropologists—and sometimes HR experts. You need professionals who understand the critical context for behavior patterns and what to do with that information.

Proper sampling setup methods can make it easier to detect new and emergent opportunities in quantitative data. There are also statistical techniques that are more responsive to changes in data, and ways to set up your data feed and your sampling. These methods could provide you with, for example, insights into patterns of attrition occurring at different stages of the coronavirus pandemic by asking questions such as: What are the patterns of security and anxiety driving the numbers?

Getting Out of the 'Data Swamp'

According to Fink, collecting data devoid of meaning or value to your particular organization puts you squarely in a "data swamp." Without the ability to understand why a pattern is emerging, you're steering blind. The refinement of data makes it valuable, as properly analyzed data doesn't just reveal what happened—it can help predict what might happen.

Most of us are good at answering the "what" questions about our organizations (operational metrics), but it's the "why" and "how" questions that provide us with valuable insights. People too often look at operational job metrics, such as time on the job, as if they are supposed to tell us something meaningful, when often they obscure the truth. You want to know if that person is prepared for a new role, but operational metrics—those "what" questions—do not hold the answer.

Smaller or newer people analytics teams tend to "get sucked into the tyranny of the urgent," according to Fink, meaning they scramble to figure out what they can do with only the data they have on hand, or what's convenient. Building a dashboard of the metrics available is an excellent way to start a conversation with your leadership, but it's far from the whole picture. You also need to figure out what information is essential to your specific organization or business, which will likely mean creating or finding new data sources. Innovative work in this area will allow you to make a real difference.

Common Things People Analytics Teams Get Wrong

First and foremost on Fink's list of common people analytics mistakes is focusing on what you have in hand, even if it isn't meaningful. Another frequently made error is not looking at the correct problem to be solved, which is typically not the one that is immediately obvious—correlation versus causation.  

Finally, though much less common, is levels of aggregation problems. Depending on how you aggregate data, it can tell different stories. For example, let's say a hospital starts outsourcing its nursing function. Because it's being outsourced, the nurses are no longer employees, so they aren't counted when looking at the hospital's level of diversity.

Given the ratio of nurses to surgeons, the hospital's overall diversity level will appear to have taken a nosedive, even if the surgeon pool has become more diverse. By looking at different levels of aggregation and who's included in a segment, you can see micropatterns that move in one direction and macropatterns that are deceptive and move in a different direction or vice versa. When we don't think deeply about the levels of aggregation and segmentation, we draw conclusions that simply aren't true, which leads to ineffective attempts at solving the problem.

Helping HR Find the Best People

Today, we have the power to conduct analyses we never could in the past. A favorite project of Fink's, with a past employer, was using artificial intelligence and natural language processing to figure out how existing skills related to one another to optimally staff new and urgent skills—thus finding "unicorns" who do jobs that maybe only a handful of people in the world could do. By dissecting the component skills that made them an expert in "X," they could find people who had the same component skills, meaning they would be able to become experts more quickly in "X" and fill that organizational need.

Fink and her team found that people who were about a 70 percent match to the requisite skill set were the best fit because they had room to grow into the role, were excited to learn, and brought new skills and thinking to the position. Someone who was a 100 percent match might quickly grow bored and move out of the role.

Too often, hiring managers look for every skill in one job candidate, which is unrealistic. "My conversations with people on this staffing issue are often, 'We can't find anyone to fill this leadership position,' and it turns out they actually need three people to fill the role because they want someone who does, say, systems engineering and benefits and statistics," Fink said. "You're not going to find that in one person. Then the question becomes how you build that coalition."  

People analytics has the power to support our organizations—if used correctly. It can help us find the right people, with the right skills at the right time; ready, willing and able to help our organizations move forward—ultimately, the end goal for us all.

Your business needs are unique. Get the answers with SHRM Enterprise Solutions, a comprehensive suite of resources customized to empower and elevate your organization.


​An organization run by AI is not a futuristic concept. Such technology is already a part of many workplaces and will continue to shape the labor market and HR. Here's how employers and employees can successfully manage generative AI and other AI-powered systems.