Data-Driven HR: Three Steps To Get Started With HR Analytics Today
As companies increasingly focus on the agile and digital transformation of HR, the best way to get started is by mapping the organizational challenges that represent priorities and pain points not just for HR, but also for the business, its leaders and its employees alike. Doing that in a data-driven way has many advantages. Just one example: a data-driven HR function could easily predict today’s talent shortages—which, in all likelihood, are only going to get worse in the majority of developed countries—five years ago, and prepare for them in time.
Reader discretion advised: this long and detailed article is not for the fainthearted! Proceed only if you have a serious interest in learning about HR analytics.
Author: Daniel Bodonyi, People Analytics & Data expert
Step #1: Identify The Problem
The best way to get started with HR analytics is to map the organizational challenges that represent priorities and pain points not just for HR, but also for the business, its leaders and its employees.
This should be done in a data-driven way, using external and internal data and both quantitative and qualitative research. There is plenty of high-quality research out there—including this piece by the Boston Consulting Group and a joint study by the Spark Institute and Quantified Company (in Hungarian)—about the trends influencing the labor market today. Perusing these can not only give you a picture of the external drivers that impact every company today; they can also help you understand the drivers that are unique to your organization, possibly because of reasons having to do with your industry or company size, or perhaps because some of your processes aren’t as optimal as they could be. Standard statistical reports of economic indicators can also provide you with useful quantitative data. Just by looking at GDP, employment and demographic trends, you can predict labor supply and demand in your country. In Hungary, for instance, talent shortages are likely to get worse in the next five years even if GDP does not grow at all. In fact, because of the country’s rapidly aging population, even an economic crisis similar in magnitude to that in 2008 might not be sufficient to restore an equilibrium-like state of labor supply and demand in Hungary.
At any rate, companies where HR was in the habit of using data and analytics could easily predict and prepare for today’s talent shortages as long as five years ago, and it’s probably fair to say that today’s labor shortages are only going to get worse in the next five years in Hungary.
Using qualitative research can also help identify your key HR challenges. Structured interviews, just like in the hiring process, can provide great help with the data-driven planning of your HR strategy. Doing at least 20-30, 45-60-minute-long structured interviews with leaders and employees across your organization can be a great addition to the quantitative data you may have collected, helping you to understand relationships of cause and effect, as well as the subjective experiences of your colleagues in different parts of the organization.
Step #2: Measure The Impact
Once you have identified your most pressing challenge or challenges, the next step is to measure the impact and calculate the ROI you expect from fixing the problem. This is crucial in getting the necessary buy-in and resources from executives to solve your challenges.
Employee turnover is an excellent example. Because of today’s talent shortages, attrition weighs heavily on many HR leaders’ minds, yet it is often hard to get executives’ buy-in for potential solutions.
Why is that the case? Partly because salaries are reported on the expenses side of income statements, so in the language of numbers—at least at first sight—every resignation appears as a cost reduction, however bad that may sound from the perspective of an HR professional and every employee. And if a resignation leads to extra work for those who stay, no problem—surmise too many leaders—, team members will just have to work harder for a while (and often without getting paid for overtime). Sales leaders, though, often represent a welcome exception from this mindset; they understand all too well how losing a sales rep can make it much more difficult for them to achieve their quota.
If HR can calculate, however, how expensive it is in terms of recruitment, training and opportunity costs to replace an employee, it becomes a lot easier to demonstrate that employee turnover is not just something unpleasant, but in many cases a significant barrier to revenue generation and growth. Our Data-Driven HR Course in November will cover this topic in more detail, but in the meantime, here is an example: at one company, a quick calculation has revealed that every resignation represents a cost of about USD 100,000 for the company. With that figure at hand, it was a whole lot easier to get executive buy-in for measures to reduce employee turnover.
However, it is equally possible for the numbers to show the opposite: that it is cheaper to replace an employee who leaves than it would be to allocate resources to retain them. An HR leader from a US-based e-commerce giant once said at a people analytics conference that it was cheaper for them to maintain a steadily high employee turnover rate on their call center teams than it would be to provide training, promotion paths and salary raises to retain them. Whether that is likely to stay that way as talent shortages increase is by no means certain, but for this company, in this specific role, employee retention is not a priority for the time being despite high attrition rates. This too could change if their competitors offer better conditions in the future, making it more difficult for the company to source talent. With a data-driven approach to HR, modeling such scenarios in real time is also possible.
Step #3: The “Five Whys,” Or Connecting The Dots
After identifying the problem and calculating ROI, the third step is to identify the root causes of the problem. In order to do so, it is helpful to use the “five whys” technique, review the relevant scientific literature (from fields like psychology, sociology and neuroscience) and international best practices, build familiarity with the available technology solutions, conduct an exploratory analysis of existing data, and build on the insights from the structured interviews done in the first step. These can all help in formulating hypotheses about the root causes and the potential solutions.
Suppose our number one goal is to reduce employee turnover. In the first two steps, we have identified the reasons why this goal is important for us—perhaps because it is growing or is higher than the industry average, it represents a high cost for the business due to rising wages, high training costs or projects lost because of insufficient capacity to take on new business, and it increases the risk of burnout among employees who stay.
In the third step, we can break down the problem and get closer to its root causes using the “five whys” technique—a simple approach that should be familiar to everyone with naturally curious small children at home.
→ The first “why,” or conducting an exploratory analysis of existing data
Not all data is created equal. The fact that employee turnover is, say, 20% doesn’t mean much on its own; we also need to understand the reasons behind it. Using existing data, we can begin to break down the problem. In what parts of the organization and among what kinds of employees does attrition represent a problem? Which sites, which units, and which roles are most affected? Do most people leave in their first year or later? Are we losing high-performing employees or underperformers? Is turnover higher among people with below-average salaries, or is attrition high in well-paid jobs as well?
Most companies do have at least some amount of data that can help answer these questions, so it is important to systematically structure that data and start connecting the dots. For example, if the majority of employee turnover is first-year turnover (as often is the case), reducing it will likely entail diving deeper into our employee selection and onboarding processes. But if we’re mostly losing people with longer tenure at the company, then we might need to look into promotion paths, performance review processes or compensation.
→ The second “why,” or reviewing qualitative research and secondary sources
Once we have localized the source of the problem—so, in the case at hand, we have identified the organizational units and employee groups where turnover is high—it can be useful to look at the qualitative information gathered in the first step through internal interviews. When looking at the groups of people most affected by employee turnover, is there anything that stands out as a potential cause in the subjective experiences of leaders and employees? How are these experiences different than the those of the people in less affected groups?
In addition, it is useful to review international best practice examples and the relevant scientific literature. Solving HR problems rarely requires reinventing the wheel, but it does require familiarity with how others have solved similar problems in the past.
Psychologists have long established, for example, that our decisions are influenced by our personalities and our motivational needs, values and preferences, and that it is primarily the latter that influence employee retention. Psychologists at Attuned.ai, for instance, have identified the 11 factors below as having an influence on employee motivation and retention:
Although each person prioritizes these factors differently, in general we can say that most employees will be motivated by a supportive environment, autonomy in making decisions and managing their time, achieving ambitious goals, appreciation and constructive feedback, financial security, varied tasks and opportunities to innovate, professional and personal development, being involved in decisions and understanding the rationale for them, stability and predictability, social relationships, and the prestige of their company or job. According to the research, the risk of voluntary turnover increases fivefold if three or more of the six factors most valued by an employee are unmet at the workplace.
→ The third “why,” or collecting data
Now that we have a general idea about where in our organization and why employee turnover might occur, it’s time to collect some data. When it comes to employee turnover, employee engagement surveys are often used for this purpose.
However, using appropriate measurement tools is critical. In the case of engagement surveys, it is important to use a scientifically validated and statistically reliable survey, designed by psychologists. It is also important not to jump to conclusions based on the raw data, but to look at the distribution instead. For example, a survey response of 3 on a scale of 5 might be an excellent result (if everyone else chooses 1 or 2). But if 80% of the respondents choose 5, a response of 4 can also indicate a problem.
Market knowledge and specialized expertise play a crucial role in this step as it is important to know what tools are available in the market and advisable to engage a specialist to validate the scientific and statistical background of the tools. Otherwise all the effort, time and money spent on collecting data may well lead to distorted results and biased decisions.
With the right tool, however, we can quickly and painlessly identify the causes of employee turnover. In high-growth organizations, for example, these causes frequently include the fact that employees perceive the organization as unstable and unpredictable, they aren’t sufficiently involved in decision-making or receive little information on the background of decisions, so they perceive them as irrational, and despite—or because of—the fast pace at work, they don’t receive enough appreciation as successes and opportunities are undercommunicated by leadership.
→ The fourth “why,” or preparing an action plan
Suppose that we have identified the causes of employee turnover using reliable measurements. For example, we have discovered that our sales reps don’t receive enough support, training and feedback in their first six months to feel confident that they can reach their targets, and that our senior developers don’t see their compensation and their professional development opportunities as attractive. We now know precisely why turnover occurs in these groups, but that is nowhere near enough to solve our problem. We also need to understand the reasons why these causes occur in the first place.
Why don’t our new sales reps receive enough support, training and feedback? It is possible that we need to improve our onboarding process, create a new, sales enablement type role, change how managers conduct 1-on-1 meetings, provide coaching and other support for managers to improve feedback culture, or outsource parts of the training because we don’t have sufficient capacity in-house.
Why don’t our senior developers enjoy more attractive opportunities? Perhaps we don’t have an adequate competency model for this area, or we have one without adequate promotion paths above a certain level of seniority, or perhaps those promotion paths are narrowly geared towards people management responsibility, which may not be attractive for every team member. Or perhaps we don’t have up-to-date, reliable data on compensation trends, and so we cannot take prompt, proactive steps to prevent our competitors from poaching our top talent.
In order to answer these questions, it is important to have a thorough knowledge of the organization and critical to involve the affected leaders and employees even if a particular issue is sensitive. Decisions made above the heads of those affected may be convenient in the short term, but are rarely optimal in the long run.
→ The fifth “why,” or preparing a communication and implementation plan
Once we have identified the actions we need to take to solve a problem, it is wise to also think about why that problem has proven difficult to solve in the past in order to identify any forces that may continue to hinder a solution in the future. Most initiatives fail at precisely this point. Seeing what the problem and its causes are is only the tip of the iceberg. Actually solving the problem is the hard part.
At this stage, our data-driven HR project becomes a change management initiative, which needs to be communicated and implemented in the right way, by the right people. The Data-Driven HR Course at the Spark Institute at IBS will present some useful models for this, too, since data also makes it a lot easier to communicate the necessity and benefits of change.
In conclusion, if you’d like to get started with HR analytics, the following are helpful to keep in mind:
- Start with the business case, break it down to functional goals, and collect data that helps you fine-tune and achieve those goals. Identify the HR measures that have the biggest impact on the success of the business. If your company wants to grow, HR can contribute by making recruitment more effective, reducing employee turnover or improving your company’s leadership culture. If the goal is to boost profitability, HR can help find ways to improve productivity, performance, efficiency, motivation, competencies, skills, and related processes.
- Data should be collected from both inside and outside the organization, using quantitative and qualitative research. Numbers and hard facts matter, especially if they are gathered using scientifically and statistically validated instruments. But qualitative information reflecting subjective experiences also plays a great role when dealing with people-related matters, so strive to enrich your data by doing interviews. It is also important to keep up-to-date with the latest research, best practices and technologies in the market.
- When collecting data, you need both “outcome” data about the results and “input” data about the potential causes. For example, if your goal is to reduce employee turnover, your outcome data will be a simple boolean “true/false” (or “yes/no”) on each person depending on whether they have left or stayed at the company, and your “input” data might include things like salary, time spent at the company, performance review data or engagement survey responses (aggregated on a team level to preserve anonymity). It is by comparing these “outcome” and “input” data points that you can analyze the differences between those who stay at the company and those who resign.
- Data-driven HR is about a lot more than data. Data can help you identify a problem and its potential solutions, but actually solving the problem is the hard part.