The human resources team plays a key role in shaping a company, so it’s important that it uses every tool in its toolbox to recruit, retain and develop the best people. That’s why more companies turn to HR analytics, the practice of collecting and analyzing data to inform human resources decisions.
What Is HR Analytics?
HR analytics is the collection and analysis of data to inform human resources decisions around recruiting, retention and other company priorities.
HR analytics can be used to identify employee trends, uncover the causes driving workplace dynamics and identify potential issues that might otherwise go undetected. Analytics can also be used to prove the effectiveness of HR policies and initiatives.
HR teams use data to make smarter decisions about everything from recruiting to employee engagement to organizational structure. This could involve a better understanding of a company’s workforce diversity or the employee-to-manager ratio at each level of the organization. They may even try to predict how many employees will resign or retire, which will be useful in planning its recruiting efforts.
In this guide, we explain the role of HR analytics, break down some examples of common metrics and use cases and discuss the various approaches that can help companies get the most useful information for their business needs.
What Is HR Analytics?
HR analytics refers to the collection and analysis of data used to help human resources teams shine a light on trends or problems they might not have otherwise noticed. It can also offer an objective source of truth for important decisions that may have been based on gut instincts or unconscious bias.
HR pros gather this data from employee surveys, performance reviews or exit interviews. They can also gain insights from the HR software systems that companies use for recruiting, payroll and HR administration.
Richard Rosenow, vice president of people analytics strategy at One Model, said all this talent data is essentially a way to listen to employees at scale.
“Instead of speaking to people one on one,” Rosenow said, “you’re going to listen through the data.”
Data analytics doesn’t replace the people-oriented work that HR professionals do, though. Although Rosenow is a data evangelist, he said HR teams still play a valuable, “almost anthropological role,” understanding workers in a way that data alone cannot.
“[HR teams] can be supplemented by data to get somewhere further, but data is not going to replace you — and you’re not going to replace data,” Rosenow said.
How to Use HR Analytics
HR analytics helps teams identify, define and solve issues within the organization, as well as improve business operations, predict future trends or assist executives in making difficult decisions. More specifically, HR teams might look to analytics to help retain employees, identify opportunities for employee training or conduct strategic workforce planning.
But how does an HR team turn its data into insights?
Michael Beygelman, executive vice president of product at Claro Analytics, suggests business leaders first determine what question they are interested in answering. From there, HR analytics teams can work backwards to find the relevant data.
That might sound simple enough, but HR teams will have some preliminary work to do before they start digging into their data.
Jeff Jolton, managing director of research and insights at Kincentric, said HR teams should start by establishing a clear definition of key measures, identifying a clear source of where the data is coming from and pressure-testing the data to make sure it is accurate and reliable.
1. Define the Business Objective
HR teams should ensure that the question they are asking is framed to give them the best possible answer — and that all departments are aligned on the definition of the metric they are looking at. HR teams might calculate cost or performance in one way, but the finance department may be calculating it a different way, Jolton said.
Calculating a company’s headcount, for example, can become complicated when one starts to consider whether they are pulling data from the start of a quarter, the end of a quarter or an average headcount over the course of the quarter. That number may also change if the company is including contractors or vendors in their headcount.
“When you really dig into any one of these metrics in HR, there’s a depth to it — almost like a quantum layer that goes really deep,” Rosenow said. “There are all of these potential opportunities for how you could define this thing, so these technologies have to define and architect those metrics to make them easy to use.”
2. Find and Extract the Data
One of the most common challenges HR teams face is finding and extracting their data. HR teams have data throughout their HR tech stack, such as their human resources information software, performance management system and applicant tracking system.
But extracting and formatting that data can often be difficult, because one software’s data may not be synchronized along the same timeframe as another platform’s data. Rosenow said many HR software platforms were not built with data analytics in mind, so it needs to be cleaned and configured to create a unified data set.
“HR data is really unique in the sense that the data from Greenhouse and the data from Workday might not play together nicely,” Rosenow said.
There are a variety of HR analytics software tools on the market that can help companies extract, model, store, analyze and deliver their HR data.
3. Analyze the Data
Once you have an accurate and reliable data set, it’s time to analyze the data and test any hypotheses you may have formed. Data scientists can interpret the data through statistical analysis and machine learning.
There are four types of analysis used in HR analytics: descriptive, diagnostic, predictive and prescriptive. The level of sophistication you use for your project will depend on what you are hoping to gain from the analytics process. These four approaches are broken down in greater detail later in the article.
4. Present and Apply Findings
HR analytics is more than just gathering and analyzing the data. Once the data has been analyzed, it is up to analytics teams to lean on their institutional knowledge about the company to describe what the data is saying, what it means for the business and what steps can be taken to address the issue or optimize operations. When presenting these findings to company leadership, data analytics teams can make their data more digestible by using data visualization tools.
“Metrics alone are merely the ‘what;’ HR analytics teams must interpret and contextualize data to form the ‘so what,’ which are the actionable next steps,” Beygelman said. “This will help HR analytics teams craft their narrative to achieve data storytelling, a powerful technique for conveying meaning with data.”
HR Analytics Examples
HR teams can look to a wide variety of metrics for answers. These metrics can span across several HR priority areas, including recruiting, employee engagement and employee performance management.
Headcount
Headcount is the number of employees that work for a company at a specified time. This metric helps HR teams engage in strategic workforce planning, which ensures the company has adequate personnel to meet its projected business demands.
Turnover
One of the most common HR metrics is turnover. High turnover can be expensive and hurt an employer’s brand. HR analytics can help companies identify if the turnover is caused by an issue in a company’s culture or recruiting process, for example. When measuring turnover, companies might look at which employees left voluntarily, which employees left involuntarily and the number of high-performing employees that left the company.
Retention Rate
When measuring turnover, companies might also look at the retention rate, which measures how many employees have stayed with the company. Companies that want to retain the institutional knowledge of their employees and avoid costly hiring processes should make an effort to listen to employee feedback, recognize employee accomplishments and invest in their employees’ professional development.
Compensation
Compensation, which includes the cost of salary, benefits and bonuses, should remain competitive with other companies to recruit and retain the brightest employees. One popular metric for analyzing compensation is compa-ratio, which compares an employee’s salary to the median compensation for similar roles within an industry.
Time-to-Hire
Time-to-hire is a key recruiting metric that measures the time between a candidate applying for a job and accepting a job offer. This metric is important in gauging the efficiency of a company’s recruiting operations. A lengthy or burdensome hiring process can also impact the candidate’s perception of the company.
Cost-Per-Hire
Cost-per-hire is another recruiting metric that analyzes the efficiency of the hiring process. It tracks all of the expenses the company incurred to fill a position, including dollars spent to post the position and the wages of the recruiters and employees who interviewed candidates.
Offer Acceptance Rate
The offer acceptance rate — the percentage of candidates that accept a job offer — is another valuable recruiting metric. If your offer acceptance rate is low, it might indicate an issue with the compensation, company culture or hiring process. It also means the company’s recruiting operation is spending an inordinate amount of time on hiring, which is an expensive process.
Diversity
Gathering data about employee demographics can help a company understand the makeup of its workforce and consider whether it has created an inclusive environment for employees from underrepresented backgrounds. If a company is lacking in diversity, it may want to take a look at potential pay gaps and consider what efforts they can take to recruit passive candidates from more diverse backgrounds.
Types of HR Analytics
There are four different types of approaches that HR professionals take with HR analytics: descriptive, diagnostic, predictive and prescriptive.
These are often described in terms of maturity. Descriptive analytics is considered the least mature or sophisticated, and prescriptive analytics is labeled the most mature or sophisticated. The type of approach used is dictated by the issue that a company’s leadership is trying to address.
1. Descriptive
Descriptive analytics help determine what is happening on a given topic, but it doesn’t explain why it is happening. If a company is looking at turnover, for example, a descriptive approach will focus on finding out the turnover rate and if there are any groups of workers that leave more often than others.
“Descriptive work is where the bulk of the value is for most people analytics teams,” Rosenow said. “We have to be able to describe our workforce, and to do that, descriptive analytics goes a long way.”
2. Diagnostic
The diagnostic approach is used to figure out why a company is experiencing a certain problem or trend. In the example of turnover, a company might use the diagnostic approach to uncover why certain groups of employees are leaving. If turnover is higher among new hires, Jolton said, the diagnostic approach might look at onboarding surveys and exit interviews to see if the employees were given an inaccurate representation of the job description or if they didn’t feel a sense of belonging.
If an inordinate number of job applicants are declining job offers at a company, the HR analytics team could use the diagnostic approach to identify what’s causing candidates to get cold feet at the last minute.
3. Predictive
Predictive analytics are used to determine what will happen in the future. It can help HR understand where the company is headed, where it would like to go and what their trajectory would look like under different hypothetical scenarios.
In the turnover example, the predictive model could be used to understand which employees are at risk of leaving. If the data shows that sales employees tend to leave when they don’t have strong connections in the office, HR can develop social programming to try to prevent those employees from leaving.
“Instead of waiting for something to happen, we’re now using insights that we’ve gained from the first two steps to better predict if they might leave — and then try to stop it before it happens,” Jolton said.
4. Prescriptive
Prescriptive analytics leverages machine learning to recommend solutions to issues identified in the other three approaches.
If a company has already learned which factors are contributing to employee turnover and predicted which types of employees are likely to leave, then HR analytics teams could use the prescriptive model to suggest interventions, like specific training programs or tweaks to its recruiting processes.
Companies working at this level of sophistication might also try scenario modeling, Jolton said, which analyzes several possible solutions to find the one that makes the biggest impact.
While the prescriptive approach is generally used after one or more of the other methods, Jolton said there are some situations where analysts might start with the prescriptive model. If an HR team is trying to decide the best type of training to offer, they might use prescriptive analytics to conduct a pilot test or a cost-benefit analysis.
Rosenow said hardly any HR analytics teams use prescriptive analytics because the decisions that HR teams make affect real people and should ultimately be made by humans instead of algorithms. Citing an IBM internal training manual from 1979, Rosenow said “a computer can never be held accountable, therefore a computer must never make a management decision.”
While many in the HR analytics field have long questioned the ethicality of automated decision-making, the practice has recently come under legal scrutiny, as evidenced by New York City’s recent adoption of Local Law 144. That law requires any company using algorithms to hire, fire or promote employees to submit their algorithms for a public audit and disclose their algorithm use to job candidates.
Jolton agreed that predictive and prescriptive models are not used as often as the descriptive or diagnostic approaches.
“The majority of the work is still going to be descriptive or diagnostic,” Jolton said. “Not everything has to become predictive or prescriptive.”
Frequently Asked Questions
What are the types of HR analytics?
The four types of HR analytics are descriptive, diagnostic, predictive and prescriptive. The level of analysis you use depends on the nature of the business problem you are trying to solve.
What is an example of HR analytics?
One common example of HR analytics is an analysis of a company’s turnover rate, which can be further broken down into voluntary and involuntary turnover. A high turnover rate can indicate a problem with the company's culture, recruiting process or professional development programs.