Data Science and Talent Management, A Match That Drives Margins

Six years ago, data science was making its first impacts on HR and hiring. Today, there is a diverse marketplace of machine learning based products. They still do not address linking ROI to individual hires or the hiring process.

Vin Vashishta | Originally Published: September 21st, 2014

Book time with me for Career Coaching or sign up for my Business Strategy for Data Scientists class.

It is well known that data science can provide strategic insights about customers but what if your customers are the business’s talent? What kind of insights can HR, a group that already benefits from access to high quality data, derive from data science? For data science to add value it needs to be focused on new, forward looking and increasingly granular insights. Just as marketing is using data science for real time, predictive and personalized, HR can too.

I like to start data driven journeys with questions. When a business looks at its people, its human talent, the obvious questions come to mind first:

  • What engages them and what disengages them?
  • How does the business attract and retain the talent it needs?
  • What is the right compensation?
  • Where do we train and what do we train on to move the organization forward?

  • What is the business interest in engagement, training, hiring, retention, and compensation? The interest revolves around value creation. HR has studies that show engaged, well trained employees produce more value than the opposite. Looking at the studies on employee value creation, that interest in value is a two-way street. Employees are more satisfied where they feel like they are maximizing their contribution to the business and receiving a fair compensation in return. It turns out that HR is in the business of understanding value; how employees create it and how they consume it. The goal is to maximize the value created while keeping the value consumed in line with the profit margins the business expects.

    Now we have an algorithm and as data scientists, we love those. We can run algorithms against datasets, and they return insights. To get the dataset we need, we must ask regression questions: how do employees create value and how does the business return value to employees? Value creation data coupled with talent management data yields the kinds of insights that drive margins higher.

    Productivity – Employee Value Creation

    How much value does any given department produce let alone an individual employee? Answering this question with increasing levels of granularity is possible with data science. The process of answering this question puts HR at the center of gathering some of the most strategically valuable insights the business will get from data science.

    It starts by looking at the business model and understanding what value the business creates. For many retailers, having this information five to ten years ago would have helped them avoid the downturn they now find themselves in. Looking at Best Buy as an example, they provide customers with a physical location to see and try out electronics. While they generate their revenues from selling electronics, their value to customers is as a showroom. Employees are trained to sell to a customer, which adds value to the business but does not add value to the customer. From an HR perspective, the hiring and training strategies are emphasizing the wrong skills and setting the business up for failure.

    Businesses have answered the value creation question when the customers’ answers line up with the company’s answers. For HR and many other groups in the business, understanding how the customer perceives the value created by the business is an essential piece of information. It is a dataset that marketing already has access to in many businesses and it becomes even more useful when it is linked to talent management data. It helps HR answer an important question; do we have the right talent to support the business model? That analysis also needs to be forward looking. The business has a three to five-year strategy and HR needs to know how value creation will change over that time to keep their strategies in lock step.

    Even in the most basic question about productivity, “what do we produce?” there’s significant value for HR. Data science provides valuable insights by predicting the cost of hiring necessary skills over the next five years and modeling training initiatives to help current employees maximize their value in a changing business. As the questions get more granular about value creation HR has an opportunity to provide insights that improve margins.

    The examination of value creation next looks at how that value is built by the business’s talent. Value stream mapping and several other tools have been used to get a high-level understanding of this process in groups with direct contact with the value creation process. What about groups that do not build products for sale; how do they create value? In many departments that do not touch the product, measuring value creation has been more art than science. From an HR perspective, advising departments on head count, skill sets, training and organizational structure only works with a solid understanding of how they create value.

    That analysis starts with a familiar process, connecting customer data with talent management data. The customers are not always obvious. I worked with a global hospitality company looking to better understand internal value creation to help them increase margins. They were struggling to create a picture of how their finance group was creating value for the business. Long data presentation short…their customers were investors, other internal teams, and government agencies. We were able to show the finance organization was running at a 20% margin in a supportable way. With the concrete understanding of how they created value they were able to restructure their activities to grow that to a 27% margin.

    HR contributed to that effort in a big way. They provided data on existing skillsets, training options versus hiring costs and helped with the restructuring. Other organizations quickly realized the value of this model. HR became a strategic partner in focusing the business on hiring, training, and restructuring strategies that drove significant improvements in operating efficiency. It was all enabled by connecting a better understanding of how the business creates value with talent management data.

    These small data wins are drivers for HR to be involved with data science initiatives. Large datasets and the resulting insights allow HR to build increasingly granular pictures of how training, hiring and retention can contribute to higher margins and prepare the business to execute on the longer-term strategy initiatives. Tightly connecting talent management data with value creation data is the key.

    Compensation – Employee Value Consumption

    Most businesses have adopted some form of market-based compensation. How does that correlate to employee value creation? Short answer, based on a lot of employee data, is it does not. The variance in value creation between two nearly identical (skills, education, and experience) employees can be massive. Instinctively we know that. However, without data science, we can only talk about the variance in general, unsupportable terms.

    Why do employees with similar skills create different levels of value for the business? Long data presentation made short…at the highest level it comes down to compensation and employee satisfaction. As studies show, higher levels of employee satisfaction are closely tied to higher levels of productivity. What we are beginning to learn is how closely tied compensation and employee satisfaction are when the definition of compensation is expanded to go beyond salary, health care and retirement.

    With data science, compensation can be viewed in a new way. Items that used to be intangibles can now be adequately quantified and related to individual employee compensation. A lot of these intangibles are well known to businesses: healthy food options, a gym, a game room, outdoor break spaces, and flexible work hours. Some of them are less well understood and I will get to that in a minute. The important connection data science reveals goes from productivity, across satisfaction, to compensation.

    Why is this important? It allows a business to create compensation strategies with measurable, supportable ROI. I worked with a mid-sized software development company looking for ways to improve productivity without increasing headcount. After running a new type of employee survey, we discovered that many of the developers were interested in a healthy lifestyle focused on diet and exercise while also believing that their work interfered with those pursuits. The business spent $120,000 that year on a gym and healthier food choices in vending machines and the cafeteria. We tracked usage statistics for the cafeteria and gym compared with productivity levels over the next six months. Long presentation short again…productivity gains saved the company fifteen full time equivalents or about $2 million over the observation period.

    Let’s talk about the areas of employee compensation that contribute to satisfaction that aren’t well understood without data science. Leadership, performance feedback and advancement are all areas known to contribute to productivity and employee satisfaction. Getting any more specific than that becomes difficult because it requires a high level of granularity. People are different, as HR well knows, and those differences lead to a variety of preferences when it comes to leadership style, getting and giving feedback as well as career path. Without data enabling that level of personalization, maximizing employee productivity is a difficult goal.

    This puts HR into a familiar role advising managers on how to get the most out of the people they lead. However, the conversation is a lot more useful when it uses individual specifics. Leadership strategies can be personalized to the individual. Teams can be built where the strengths of individual leaders are matched with the preferences of team members. This methodology also extends to the hiring process. Fitting candidates by these preferences to teams with similar preferences and leadership strengths simplifies the selection process while improving the candidate pool. Data shows teams aligned this way are as much as ten times more productive than teams formed with current best practices.

    When it comes to compensation, the bottom line is data science enables HR to create a value driven compensation strategy with measurable, supportable ROI. It allows HR to see past dollars as the metric of compensation and start using satisfaction as a measure of compensation. It also allows compensation to be viewed holistically including elements that without data science become difficult to quantify. Once HR has this information specific talent management strategies can be built to improve productivity. That is a big strategic win for HR and for the business as a whole.

    Looking at People as Individuals

    The future of engagement is personalization. Marketing understands that customers are expecting increasing levels of personalization. Employees, especially top talent, are also coming to expect employers to be able to interact with them on a personal level. They are looking at compensation packages as more than money. Over half of all men and over 70% of women would turn down a higher salaried new position if it meant they believed they would not get along with their new co-workers. They expect companies to put them to work where their contributions are most valuable to the business.

    For all these reasons, HR needs access to the insights derived from advanced analytics and large datasets. HR also has a high-quality dataset that can create significant value when combined with data from customers and other departments. That combination drives margins and productivity. Companies that do not develop a data enabled HR team will find it difficult to compete in the next three to five years.