Despite all the conversation about measurement, metrics, and analytics, many human resources and learning professionals are still skittish when it comes to trusting data. Their lack of trust is not unwarranted; while, on one hand, it’s hard to dispute quantitative data, on the other hand, quantitative data do not infer we throw caution to the wind.
For example, managers in organizations track a variety of people metrics to monitor how efficiently they run their business. Revenue per employee is one such metric and one that some suggest is the only performance goal an organization needs for an HR leader. The higher the revenue per employee the better the organization is doing in terms of optimizing its employee resources.
The formula is simply Revenue / Number of Employees.
To improve revenue per employee, an organization can either generate more revenue or reduce the number of employees. It’s simple math. According to a recent Wall Street Journal article, the average revenue per employee for the largest U.S. companies has climbed 22% since 2003. While on the surface, one could surmise that organizations are more productive than in the past, this summation is not necessarily accurate. This 22% boost is due in part to a decrease in full-time employees and increase in contract labor – contract labor doesn’t fit into the equation. So, while the ratio of revenue to employee may be improving (simple math), the reality may be stagnant revenue, fewer full time employees, and more contract workers.
1. Only trust numbers; never trust numbers.
HR and learning professionals are not the only people skittish about numbers (though only 30% of professionals believe the human capital analytics function has credibility, according to a study by i4cp and ROI Institute). Any mathematician will tell you: only trust numbers; never trust numbers.
Why the conflicting advice?
The output of math is true:
1 + 1 = 2; 6 ÷ 2 = 3. In fact, mathematics is known to be the only proof due to its output being final and binary. This means that the output of math is what is – it’s not going to change.
But numbers can be wrong and misleading. Incorrect numbers occur because of lying, cheating, poor data quality, and miscalculations. Misleading numbers can occur because of inputs and context.
Trusting data takes effort. Here are a few tips that may help you.
2. Look behind the numbers.
There is always more than one way to measure something. So, get to know the inputs of a metric. The more you know about what goes into a number, the better you can judge its quality and usefulness.
3. Embrace imperfection in data.
Recognize what data tell you and what they do not. Acknowledge the uncertainty, and based on that uncertainty, consider what is probable rather than what is absolute.
4. Check the math.
While data are imperfect, math is either right or wrong. Typically, when it is incorrect, it is due to a simple oversight in calculation. Double check the math – or have someone else do it. There’s nothing worse than making a major funding decision based on a bad calculation.
5. Use good judgment.
Mathematical models can produce meaningless information. A classic example is that describing the statistically significant relationship between the number of pirates and global warming. As shown below, the number of pirates in the world have decreased as global warming has gotten worse. Does this mean we need to increase the number of pirates to stop global warming?
6. Put the numbers into context.
A metric is only meaningful when it is put into context. Before deciding on a plan of action, clarify the business question the metric represents, the inputs and analysis that lead to it, the sources of the inputs, and how it compares to target performance.
By knowing what goes into a number and embracing the inherent flaws in data, you will learn to trust your data enough to make decisions that help move your organization forward.
To help your organization develop trust in data, work with your analytics team to offer a course in the meaning behind the numbers. Learn about the metrics that matter, the analytics process, and how data are influencing decisions.
Patti Phillips, Ph.D., is the president and CEO of ROI Institute and Chair of i4cp’s People Analytics Board.