Practically every business shares the same biggest cost – employees. This makes sense – even in this age of robots and computers, human talent is behind everything that a company does. People are the source of innovation, growth, and competitive edge for every company.
Given this importance, it’s a bit strange that data science is only beginning to look inward at the workforce. We have measured the consumer behavior from every angle. We can quote the lifetime value to our customers to three decimal points, though we don’t really know them. Our employee relationships are deeper, longer-term, stickier, and more laden with potential value than customers in almost every industry.
But, most hiring and employee development happens by intuition or chance. Long-term workforce planning is only done at a very high level for very common roles. Rules of thumb and industry "benchmarks" from magazine articles dominate employee strategy. There are more rigorous and methodical methods available.
From a GAAP accounting perspective, most employee expenditures are considered to be costs. Employees are not subject to depreciation, as are machines, for example. This isn’t changing anytime soon.
But from a management point of view, employees are more like a portfolio of assets – with interlocking strengths, weaknesses, and capacities. If you run a call center, your staff is your primary tool for processing calls, arguably more important than the phone switch. Likewise for a software development group or a marketing department and their tasks.
In this sense, employees are key assets, much like machines, for getting work done. If you hire or develop a more efficient and long-lasting employee, production will go up. We encourage employers to not only value their workers as human beings, but also to think of people as productive units that can be intelligently optimized for targeted outcomes.
In this chapter we present a powerful first step to advanced employee analytics. We will walk through a cohesive framework for measuring the cost, performance and attrition of a workforce.
These three key metrics (cost, performance, attrition) combine in interesting ways to inform decision-making. Most importantly, the metrics are a quantitative baseline for predictive analytics exercises. Only with these metrics will we be able to learn how to apply predictive models or whether the models are improving operations.
Note that each of these three metrics are not a single number, but a series of values across the lifetime of an employee. Technically they are a time series or vector, but we will use the terms "metric" and "curve" interchangeably to describe them.
One notable takeaway of this chapter is a lifetime value of an employee in a role. The lifetime value of a customer is often defined as "a prediction of the net profit attributed to the entire future relationship with a customer."
Our calculations of Employee Lifetime Value (LTV) refer to "a prediction of the net profit attributed to an employee through their tenure in a given role." If someone changes roles, say from Customer Service III to Internal Sales IV, a new lifetime value will apply.
We limit our analysis to one role at a time – say Underwriter I, or Teller II, or Customer Service III. We don’t attempt to predict an employee’s promotion path or their ultimate journey through the company.
Lifetime Value is probability-weighted by the risk of attrition from the role. Because it is risk-weighted, the LTV is an excellent roll-up number to compare programs or scenarios, with a direct tie to the bottom line.
We address employee cost, performance, attrition and lifetime value to bring the practice of Human Resources into the information age. Only with these metrics will the actual value and dynamics of our "Human Resources" be known. We imagine a future where enterprise software dashboards commonly report this information, and the curves are routinely used in planning. With this kind of information innovation, business operations will be able to routinely measure and apply productive gains from advanced analytics.
Executives will learn key workforce measurement concepts that will enable a new level of business intelligence. With these key metrics in focus, executives will be able to lead predictive analytics efforts throughout the enterprise.
Human Resources, Recruiting, and Staffing Professionals will learn to measure what happens to new hires after they leave the HR funnel and enter the workplace. Intelligent feedback from workplace performance/attrition can effectively inform hiring efforts to be smarter and more targeted.
Line of Business Managers, such as Sales Operations or Call Center Managers, will learn to format their employee operations into enterprise-relevant information.
Data Scientists, Analysts, and IT professionals will learn the business context for their analysis. These analysts can apply the knowledge to identify information sources and formats for dashboards and predictive analytics.
First, it is only meaningful to evaluate one role at a time, one company at a time. The curves and dynamics for Accounting will be very different than that for Inside Sales. Likewise, "industry benchmarks" are next to useless – companies differ, regions are different, and enterprises evolve over time. It is easy enough to gather this information for your own company’s roles, and we suggest investing the time to simply do so.
Some roles have more volume and size than others. A Call Center or Underwriter role will have plenty of data for great accuracy. Executive leadership is a small sample with less turnover – not as useful for analytics.
Often this exercise comes about in response to attrition or training issues, which manifest in the first year or two. The simplest of all is an entry-level position that automatically promotes after a year or two.
These short-term cases are easier to calculate than long-term employees. Beyond a few years, cost and performance factors get more complicated with raises, equity, inflation, and the time value of money. Long-term employees also vary in performance patterns – some continue to learn, while others coast or "check out."
Three common-sense questions underlie our three metrics:
These calculations are usually done at a high level, aggregating costs and performance for everyone in a single role. More advanced approaches seek out clusters or individual patterns across thousands of employees. These "big data" approaches are more useful for transaction and revenue-related roles.
Figure 1. The Three Curves
The Cost Curve tracks how much money is spent on an employee in the role, over time. It is like a daily log of costs for a new employee, from day 1. After a flurry of recruiting, orienting, training, and on-boarding costs, the expenses typically level out as salary and infrastructure.
The Performance Curve estimates the contribution an employee makes to the company, starting at day 1. Typically until training and orientation are complete, that number is zero. Employees typically "ramp up" to productivity after weeks, months, or years. After ramp-up, that level may plateau, increase gradually, or even bow downwards after years. The ultimate level of contribution can be calculated directly for some roles, estimated for others.
The Attrition Curve shows the probability of an employee being in the role at different points of tenure. On the first day, that number is close to 100%. This number is the flip side of turnover – if a role has 40% annual turnover, there is a 60% chance of being on the job in a year. The full curve is easily calculated from the HR System of Record, and is a powerful tool.
The three in combination are exceedingly powerful.
In Part 2 of this blog, Pasha will dive more deeply into these 3 curves and show how they affect an organization’s ability to predict performance, turnover and employee lifetime value.
Pasha Roberts is chief scientist at Talent Analytics Corp., a company that uses data science to model and optimize employee performance in areas such as call center staff, sales organizations and analytics professionals. He wrote the first implementation of the company’s software over a decade ago and continues to drive new features and platforms for the company. He holds a bachelor’s degree in economics and Russian studies from The College of William and Mary, and a master of science degree in financial engineering from the MIT Sloan School of Management.
People Analytics in the Era of Big Data: Changing the Way You Attract, Acquire, Develop, and Retain Talent Hardcover – April 25, 2016
by Jean Paul Isson (Author), Jesse S. Harriott (Author), Jac Fitz-enz (Foreword)