The Rise of Talent Analytics
Jeremy Shapiro and Thomas H. Davenport
WHEN A SENIOR MANAGER suddenly announces her resignation at your organization, what typically happens next? A panicked call to human resources? Emergency meetings on how to handle the transition? A hasty phone call to your favorite headhunter on speed dial? Some organizations still panic when a senior leader leaves. But many do not because through analytics, they know more about their people than their competitors do. These companies not only know exactly how much the departed senior leader was contributing—they also understand what internal relationships the leader depended upon to get the job done, what five other people could do the leader’s job just as well as she could, and the value of the job at the company. These companies are able to sustain the good work of the departed employee and increase revenue without missing a beat.
Companies like Lowe’s, Comcast, Google, and Caesars Entertainment have adopted talent analytics. They use data about their employees to find new insights that link people decisions to organizational performance. Some companies have transformed themselves through the use of these approaches. Many, like Caesars (formerly Harrah’s) Entertainment, are in service-intensive businesses.
TALENT ANALYTICS AT CAESARS ENTERTAINMENT
Caesars, the global operator of more than fifty casinos, is a well-established analytical competitor. The company employs analytics across the business in areas including marketing, pricing, and employee/customer interactions. It turns out that in the controlled environment of a casino, where privacy is assumed to be almost nonexistent, the thousands of cameras installed to stop fraud and theft can also observe how customers react in different situations. Caesars paired industrial psychologists and casino veterans to examine thousands of hours of footage from the casino floor. What did they find? That customers, win or lose, stay longer in the casino if they are made to feel welcome there by casino employees. And apparently, as customers, we need this reinforcement about every ten minutes.
Every interaction with a customer is under scrutiny at Caesars. Take, for example, the length of the hotel check-in line and staffing for it. On a Friday night, how much does a long wait impact revenue? Friday is a popular night to check in since people are starting the weekend. Each moment someone is standing in line, he is not doing what he is eager to do, whether it’s shopping, gambling, or seeing a show. Caesars was interested in learning more about what the optimal staffing level would be for the hotel front desk and the impact on revenue from optimizing staffing levels. To find out, Caesars turned to its data. Cameras could capture wait times, and the company has the ability to see how customers interact with different games and attractions through customer loyalty cards that patrons insert into slots to get “comped” rewards from the casino later.
Crossing data from camera observations and loyalty cards with staffing plans and financial measures, Caesars changed the way it recruited, staffed, and trained employees. The company focused specifically on fifteen key positions with a direct impact on customer loyalty and revenue. For these positions, Caesars could define the impact of optimal staffing on its profit and loss statement. Caesars made full use of its data to change the way it deployed its talent, measured the impact, and optimized operations for the right customer experience. Ultimately, this lifted the company’s top and bottom lines.
ANALYTICS AND THE SERVICE/PROFIT CHAIN
Organizations are beginning to make connections between people performance, customer satisfaction, and profitability. One common approach to that relationship is called the service-profit chain, which specifies that as employee engagement and satisfaction rise, customer satisfaction and spending also increase. One reason that Caesars exploits this relationship is that Gary Loveman, the company’s CEO, was a coauthor of the published research on the service/profit chain.1
However, you don’t have to coauthor the research to employ the ideas. A variety of other companies have adopted the approach, including retailers like Sears and Lowe’s. Sears dramatically improved its performance in the mid-1990s through what it called the “employee-customer-profit model.”2
While most organizations run some type of employee opinion survey, home improvement giant Lowe’s has studied the exact impact of employee engagement at its stores. It turns out that the difference between its most highly engaged and lowest engaged store, all other differences boiled out, is $1 million a year in sales. And what does it take to achieve that $1 million? Better leaders who know how to engage employees, great hiring practices, and better functioning teams.3 Other companies, including Limited Brands, Best Buy, and General Mills, have done the same types of analysis—with similar benefits.
THE OVERALL RISE OF ANALYTICS
Talent analytics are just one example of the increasingly widespread application of metrics and analytics to key business processes. From finance to supply chain management to marketing, managers are increasingly making decisions on the basis of data and analysis rather than intuition or experience alone. Data of various kinds—from enterprise systems transaction data to Internet data to sensors of all kinds—have proliferated throughout organizations. Virtually every industry has experienced this transformation, and even those once reliant on intuition—the entertainment and retail industries are good examples—have begun to embrace more science-based decisions.
After mastering the transactions and getting data in order, executives begin to explore how they can make better decisions using the data. This is exactly the state of human resources and talent management. Most organizations now have basic HR transaction systems and databases in place. The function is poised on the edge of a dramatic change in decision making about people.
OF WHAT ARE TALENT ANALYTICS CAPABLE?
It turns out that converting data about your talent into actionable insights can have big impacts no matter how advanced your organization’s capabilities are. To illustrate, consider the hierarchy of talent analytics that ranges from HR facts at its most elemental up to sophisticated talent supply chain analyses, as shown in Figure 1.
FIGURE 1. THE HIERARCHY OF TALENT ANALYTICS.
Human Capital Facts
The foundation of talent analytics are human capital facts—a single “version of the truth” regarding individual performance and enterprise-level data such as head-count, performance levels for each employee, contingent labor use, turnover, and recruiting. Companies need to think carefully about what facts they need to get that baseline.
Particular metrics can be identified through general logic (the employee engagement or satisfaction score, for example, seems an obvious metric for many companies) or specific analyses that show that a metric is important to some other performance measure. For example, Caesars found through research that a key factor in customer satisfaction was the extent to which front-line service employees smiled at customers (which, like customer behaviors, could be measured through ceiling cameras). The company doesn’t measure smiles for every employee every day, but it did embark upon a program to remind employees to smile at the start of every work shift.
Google, whose strategy relies heavily on hiring high-quality employees, devotes considerable effort to determining who the best people actually are, both before and after they are hired. Employees are scored on twenty-five performance metrics, from how frequently they have hosted “tech talks” to the variance of the assessments given potential recruits after interviews. For more detailed analyses, Google gathers up to 200 “biodata” metrics on samples of employees to determine what factors might be correlated with performance.
For some companies, there may be one or two unique data points that indicate overall health of the employee relationship. For example, JetBlue created a new employee satisfaction metric around the willingness of employees to recommend the company as a place to work. This “crewmember net promoter score”—modeled after the metric used for customer satisfaction—has been used to understand the impact of compensation changes, to predict operational and financial performance, to evaluate the impact of employee training programs, and as a factor in executive bonuses. JetBlue also asks employees if they’d recommend the company each year on their hiring date, so effectively the company has a 1/12 sample of all employees’ opinions every month, which is used to monitor employee engagement.
JetBlue and other successful organizations are completely transparent with end users on the common definitions of these facts; any manager or employee can see how the data was collected, what formulas are being used, and importantly, why the data matters to the operation. For example, Caesars provides management-friendly documentation in its HR scorecard to ensure that all readers understand how “facts” are created and why they matter to their daily management responsibilities.
Analytical HR
Analytical HR takes metrics to the next step—not simply reporting on past outcomes but analyzing the relationships among HR and other variables. Companies displaying analytical HR approaches would also be likely to group together HR data or segment the data to gain insights into specific departments or disciplines—allowing a manager, for example, to see if staff turnover intervention is needed within the East Coast sales team but not the West Coast team. Analytical HR integrates individual performance data such as personal achievement of key result areas with HR process metrics (such as cost and time) and outcomes metrics (engagement, retention, and nonfinancial business outcomes).
For example, Lockheed Martin built an employee performance management system with the ultimate goal of measuring each employee’s performance and linking it to organizational objectives. The process is continuous throughout the year and uses an automated system to collect timely employee review data. This data can then be compared to learning management information. Lockheed Martin not only understands who has undergone formal training in specific areas, but it then identifies and monitors areas for improvement. Through these analytical tools, the company more clearly understands employee performance and can identify its highest performers and highest potentials for special programs.
Human Capital Investment Analysis
Human capital investment analysis allows an organization to understand what knowledge, skills, abilities, and connections matter most to the business and allows decision makers to understand what kind of recruiting or learning and development investments create the biggest payoff. Companies practicing this approach analyze their talent DNA to replicate and grow to their strategic objective.
For example, General Mills had been running an employee opinion survey for years. At the end of the survey, the company did what most organizations do: thanked employees for being forthright, pointed out areas the organization can improve upon, and handed managers an analysis of their own populations along with employee comments. Then, the company took a second look at its analysis regarding an employee’s willingness to put in an “extra mile” effort at work. It turns out that the employees of the top managers were twice as willing to go the extra mile at work compared with the employees of the second-best rated managers. For General Mills, excellence in management pays off in big ways. The company told everyone about what had been learned and embedded the most effective managerial behaviors into the executive learning program, which was named the “Great Managers” program. The company literally intended to understand what makes a great manager at General Mills and embed those qualities into as many managers as it could.
Workforce Forecasts
Workforce forecasts may use turnover, succession planning, and business opportunity data to create pipeline analyses identifying shortages and excesses of key capabilities long before they happen. The analyses can work off internal plans and forecasts to ensure alignment with the enterprise.
Valero Energy, an energy company with the largest number of oil refineries in the United States, is one of the most aggressive adopters of this idea. The company needed to grow in an environment of critical skills shortages within its industry. Its model analyzes historical employee records for turnover patterns and creates three-year forecasts by position, location, level, division, department, and salary—and relates them to talent supply for key job categories. The model also incorporates talent needs for future capital projects and new systems, services, and projects.
Talent Value Model
Talent value models seek to answer questions like, “Why do employees choose to stay with a company?” One way to answer that question is by using analytics to calculate what employees value most, and by using that data to create a model that will boost retention rates. For example, a value model can help managers design a personalized set of performance incentives or to assess whether they should match a competitor’s employment offer to retain an employee or decide when it is time to promote an employee.
Google, for example, uses employee performance data to determine the most appropriate intervention to help high- and low-performing employees succeed. Laszlo Bock, vice president of people operations at Google, told us, “We don’t use performance data to look at the averages, but to monitor the highest and lowest performers of the distribution curve. The 5 percent of our lowest performers we actively try to help. We know we’ve hired talented people, and we genuinely want to help them succeed.” The hypothesis was that many of these individuals might be misplaced or poorly managed. A detailed analysis of them found that this was often the case, and many difficult situations have been addressed by understanding what each individual’s needs and values are. Google also used similar approaches to learn which employees were likely to leave the company.
Talent Supply Chains
Finally, with talent supply chains, companies can quickly make decisions about any number of talent demands—from optimal hiring techniques based on up-to-the-minute data to workforce forecasts integrating new data in real time. This is at the top of the ladder because it requires high-quality human capital data, effective analytics, and the integration of broad talent management and other organizational processes. We’re not sure that any organization has yet mastered this level of talent analytics, but many of those we’ve already described are moving toward it.
GETTING STARTED WITH TALENT ANALYTICS
The DELTA model is an approach for understanding and prioritizing how to build analytical capability within an organization. It, or a similar model, is critical for organizations wishing to build their talent analytics. DELTA stands for Data, Enterprise Thinking, Leadership, Targets, and Analysts.4 Each of these key components of an analytical orientation is described below.
Data
Without acceptably clean data, analysis is impossible. Identifying what data should be captured in a defined and repeatable way is core to competing on talent analytics. If you dislike your organization’s data quality, you are not alone. Many people are frozen into inaction by data problems. Consider, however, the vast quantities of people data we do have access to that is useful. People generate huge quantities of data simply by doing their jobs and through the basic systems that support them. And with modern HR transaction systems and databases in place within most large corporations, there is better data available than ever before to feed analytics initiatives. Data is important, but equally important is the ability to understand what raw measurements are critical to organizational results. Think back to Caesars; measuring the number of smiles a customer receives is data, too.
The data category also includes analytical technology. It is necessary to have technology to solve business problems using talent analytics. However, the bar has never been lower to acquire the right technology for the job. Options range from hosted business intelligence solutions that will start your journey in talent analytics within a week to enterprise systems that are more powerful than ever. You may find that the greater challenge is capturing the data you want—not reporting on it.
Enterprise Thinking
Silos are the enemy of analytics. The most effective analytics cut across HR, finance, marketing, and customer service in order to find the nugget of insight that could create new revenue or unlock the potential of a team. It’s common to combine something as mundane as employee transfer data with customer satisfaction, gross sales, and employee performance data to find departments that are creating pools of your best salespeople and which have a negative impact on customer satisfaction. Some organizations have also suggested to us that the reason they do not pursue service/profit chain analysis is because there is no one function within the organization that controls all the necessary data.
Leadership
Don’t projects always run more easily when they are supported from up above? This holds particularly true for insights regarding talent analytics. Managers are prone to trust their gut when it comes to managing their own people. The truth is, unsurprisingly, that our instincts often mislead us. For example, while most managers believe they are exceptional interviewers, 50 percent of hires by managers using pure gut instincts do not quite work out according to executive hiring experts Geoff Smart and Randy Street.5 That fact does not tarnish a manager’s belief that when it comes to people matters, the gut rules.
When leadership (within and outside the HR function) believes in the power of data, you can create a better result—and the organization can dissolve old assumptions about how to make talent decisions. A clear, actionable set of analyses that help grow the top line or yield savings will do the trick nicely.
Targets
A key aspect of succeeding with talent analytics is a clear sense of an organization’s targets or objectives for analytics. Are you attempting to improve recruiting, retention, or customer service? It is difficult to do everything at once. Google, for example, had a strong target on recruiting when it was hiring 100 people a week. When its growth subsided somewhat, it switched its primary target to retention.
Analysts
What kind of people do you need to create talent analytics? Do they all need to be math geniuses? In short, no. The most successful companies certainly do have their fair share of statisticians, business intelligence gurus, visualization experts, and even Hadoop-capable data scientists. However, we would trade most of those skills for a team of people who understood their operations, could think creatively, and told a good story with data.
A FINAL THOUGHT ON TALENT ANALYTICS
Great HR leaders never worry about having a seat at the table in their organizations. Increasingly, the stack of paper or iPad they bring to meetings is data. But the data they are bringing are not plain old headcount data or other traditional HR information. They are bringing insights to leaders to help improve operations, fuel innovation, or execute an organizational goal. Talent analytics have helped many leading companies to create a quantifiable difference in their management of people and the execution of their strategies, and it can for your organization as well.
References
1. J. Heskett, T. O. Jones, G. Loveman, E. Sasser, and L. Schlesinger, “Putting the Service-Profit Chain to Work,” Harvard Business Review (March–April 1994).
2. Anthony J. Rucci, Steven T. Kirn, and Richard P. Quinn, “The Employee-Customer-Profit Chain at Sears,” Harvard Business Review (January–February 1998).
3. C. Coco, F. Jamison, and H. Black, “Connecting People Investments and Business Outcomes at Lowe’s: Using Value Linkage Analytics to Link Employee Engagement to Business Performance,” People and Strategy, 34 (2011), p. 2.
4. T. H. Davenport, J. Harris, and R. Morison, Analytics at Work (Boston: Harvard Business Press, 2010).
5. G. Smart and R Street, Who: The A Method for Hiring (New York: Ballantine, 2008).
Jeremy Shapiro is an HR executive at Morgan Stanley, responsible for its data and analytics function. He speaks and writes frequently on HR analysis, and is the co-author of Harvard Business Review’s October, 2010 cover story, “Competing on Talent Analytics” with Tom Davenport and Jeanne Harris. Jeremy serves as the Metrics and Measures taskforce chair for the Society of Human Resource Management, developing HR metrics standards certified by the American National Standards Institute.
Prior to Morgan Stanley, Jeremy worked at the Omnicom Group consulting with organizations such as GE, Motorola, KPMG, and BAE Systems. He was an early influencer in developing online recruiting practices in the 1990s, and co-founded Hodes iQ, a leading online applicant management system in 1999.
Jeremy is a teacher within Cornell ILRs continuing education program on analytics. He holds an M.S. in information systems from NYU’s Stern School of Business and a degree in economics and history from Rutgers University
Tom Davenport is a visiting professor at Harvard Business School. He also serves as the President’s Distinguished Professor of Information Technology and Management at Babson College, is co-founder and research director of the International Institute for Analytics, and is a senior adviser to Deloitte Analytics. He pioneered the concept of “competing on analytics” with his best-selling 2006 Harvard Business Review article (and 2007 book). His most recent book is Analytics at Work: Smarter Decisions, Better Results, with Jeanne Harris and Bob Morison. He has written, or edited, twelve other books, and has written over 100 articles for such publications as Harvard Business Review, Sloan Management Review, the Financial Times, and many other publications. In 2003 he was named one of the world’s “Top 25 Consultants” by Consulting magazine. In 2005 Optimize magazine’s readers named him among the top three business and technology analysts in the world. In 2007 and 2008 he was named one of the most 100 influential people in the information technology industry by Ziff-Davis magazines.