- Eyeballed the jar for a few seconds and picked a number that was “good enough”?
- Used the guesses listed on the paper as a guide and chose something in the middle?
- Had a “process” such as counting how many beans covered the bottom of the jar then multiplying that by the number of layers in the jar?
Our brains are great at working out how important a decision is and then deciding how much effort needs to go into making sure that decision is accurate. Sometimes, we don’t have the time to painstakingly review all the data at hand to find patterns and make informed decisions on what to do next. Sometimes we do. And sometimes, we have HR analytics software to handle the analysis for us.
In the above scenario, the method you chose was probably dependent on how much you like jelly beans. But, what if we were analyzing something more valuable? Something like the productivity and performance of your workforce against competitors in an unpredictable market. This is where the importance of accurate predictive HR analytics comes in.
RELATED: Ultimate guide to HR analytics
Why should I care about predictive HR analytics?
Predictive HR analytics is a methodical approach to decision making that can use reliable past data from your employee feedback, training and development sessions, eNPS (and so on) to predict future outcomes that your business cares about. Sounds great, right? Here’s why the importance of HR analytics is so prevalent right now:
- Your competitors use predictive HR analytics – or they’re about to:
- This is a rapidly growing approach to strategic people management that can make or break your bottom line which is why it’s an important method to factor into your strategy if you wish to remain competitive.
- Wayne Gretzky once said, “I skate to where the puck is going to be”. When we talk about this in the context of predictive HR analytics, as a business, it gives you the opportunity to stay ahead of the game; you’re the only one who truly knows where the “puck” is going to be and by the time your competitors catch up, you’ll be on to the next goal.
- Predictive HR analytics makes decision making efficient so HR can become a true business partner:
- Your Finance, Marketing and Sales teams can probably all make data-based predictions, whether it’s knowing certain types of advertising will likely gain more engagement or understanding what time of day to call prospects for a better chance of success with a sale. With HR analytics, you’re able to add value from a HR perspective to your overall business objectives with clear data to support your decisions and budgeting.
RELATED: 4 Ways to use HR analytics to build a better workplace
How to use Predictive HR analytics
If you’re new to the world of collecting, analyzing and using data, it’s easy to feel overwhelmed and unsure where to start. We believe the best place to begin is by basing your process around these three key ingredients for success with predictive HR analytics:
- Get the data
- Understand what it means
- Act on it
That’s basically it… (Of course, you can go deeper, and the more you dig, the better your results) but if you’re just starting out, we recommend choosing one question and following it through with these steps. As an example, let’s look at “attrition” as a factor a business might want to understand.
1. Get the data
This step can either be easy or very difficult depending on if and how you’ve collated your organization’s and employee’s information up to this point. Most businesses keep track of at least some of the basic HR data that can be used for predictive analytics:
Let’s dive deeper into some specific HR analytics examples, along with the attrition data you’ll need to analyze them.
Starting with the basics
You can kick off predictive analytics with general attrition data. Most payroll records will capture the tenure of your workforce, allowing you to do some basic analysis such as finding out averages and changes over time. This lets you answer questions like:
- Is our turnover rate rising, falling or staying the same over the years?
- Are there certain times of year where attrition peaks?
Taking it to the next level
You can use a more comprehensive set of data to understand specific issues such as why people are leaving. Of course, the easiest way to do this is to ask them: this is why it’s a great idea to have an exit interview or survey as part of your offboarding process (if you can automate it; even better!). However, if you combine collated data with other information you have on peoples’ roles, you can start to answer questions like:
- Is there a pattern in the reasons people say they’re leaving? (e.g. for other roles, money, greater development opportunities)
- Are attrition rates higher in certain teams, departments or locations?
- Are people more likely to leave after they’ve been here for a certain period of time?
- Are people with lower satisfaction scores or performance ratings more likely to quit?
- Is employee sentiment (positive or negative language used in feedback) an indicator of the likelihood of someone staying or going?
intelli tips: Here are points that will help you when you’re collecting data:
- Don’t get too hung up on data collection as a task – If you’re handling your processes right at an employee level with continuous feedback, performance tracking and keeping your records up to date, the data should be created almost as a byproduct of existing activities, rather than an extra task.
- More data is better – This goes for the amount of years’ worth of data and how broadly the data spans. Start capturing it now to get as much information as you can.
- Be patient; collecting data takes time – Even more so if you weren’t previously tracking and analyzing this kind of information.
- Good data is crucial – You’re about to use this to make some potentially big business decisions. It’s important that it’s accurate, or your decisions will be way off.
- Make it easy; keep everything in one place – This is where good HR software really shines; it’s built for this purpose and will make accessing data much easier than spreadsheets.
RELATED: Complete guide to employee turnover and attrition
You don’t have to be a data scientist to conduct a successful predictive HR analysis, but there are two concepts that will help you stay on track:
- Statistically significant = A result that’s very unlikely to happen by chance. This is important because you want to feel confident that the relationship is real and reliable enough to base business decisions on it.
- Correlation vs causation = Just because two things often occur together (correlation) does not mean one makes the other happen (causation).
- The famous example describes ice cream sales going up each year at the same time as incidences of shark attacks rise.
- It’s tempting to conclude that sharks must find ice cream just as tasty as we do, but these two events are actually both explained by the season.
- Warmer temperatures in summer lead to more people eating ice cream and more people swimming in the ocean.
2. Use the insights to make good decisions
Now that you’ve collated your data, it’s time to turn it into something useful: insights. Like we said, you don’t have to be a data scientist to use predictive analytics; good HR software will handle a lot of the calculations for you. Don’t worry, you’re still a vital part of the process! While you don’t have to handle the data crunching, interpreting the output and making business decisions based on these insights is still up to you.
Using our attrition example, your decision-making process might look like this:
- You identify patterns in the last three years of attrition data that show:
- Turnover is 16% higher in teams with more than six people (broad span of control).
- 32% of turnover happens in the first six months of a person starting employment.
- 55% of employees identified as “high performers” said that “development opportunities” was their main reason for leaving.
- 43% of employees with satisfaction score averages of 5/10 or below left the business within eight months.
intelli tips: Before you start making any decisions, remember that these are just patterns in the data. Like any statistical analysis, predictive HR analytics has some rules to help ensure the decisions you make are based on solid foundations. You want to be confident that these patterns are statistically significant (worth paying attention to), and to understand the relationship between these variables.
- You decide if these patterns are worth doing something about:
- Depending on the size of the business and operational constraints, a slightly higher employee turnover rate across your bigger teams might not be an issue. If every operational team has seven people in it but the office team is a four-person group that includes the business owner, the difference in turnover is completely reasonable.
- Early turnover can be an indicator of a recruitment process that isn’t finding the right candidates or an onboarding experience that isn’t working effectively. If this is a significant pattern, these are two places to start investigating.
- If your high performers are consistently leaving for “better opportunities”, this is a chance to get proactive and make your organization the better opportunity. What development, recognition, compensation and remuneration opportunities could you create, and what could that save you in turnover each year?
- Satisfaction can be a tricky indicator. Are your people leaving because they’re not happy in their roles, or is something causing them to be both unhappy and to want to leave? This is where you as the expert on your workforce can dig deeper into your sentiment analysis data and understand the root cause.
RELATED: How to use people analytics for decision making
Spotlight on survival analysis
Survival analysis, while it may sound dramatic, is a powerful predictive analytics tool that draws on all of the information you have about your employees to:
- Identify the factors that have the biggest impact on whether they stay or leave the business.
- Compare everyone who has ever left the business against the people who are still there and calculate the biggest predictor of whether they stay or go.
- Take the noise (ups and downs) out of the picture to purely show the odds of a person leaving based on a specific difference. This is very useful in unpredictable times like right now as “The Great Resignation” continues to affect HR managers and teams.
intelli tips: These considerations may help you decide on whether or not the patterns in your data are valuable to your business:
- You want to stop your high performers from leaving where possible. These predictors need to be clear though: What exactly is a high performer and what is their value to the business?
- Not all patterns require action.
- Use your experience and understanding to give insight into the data.
- Don’t try to make your data fit the answer you were expecting.
- Watch out for discrimination and AI bias. Just like with Machine Learning, there’s a risk of reinforcing biases that already exist (i.e. all successful managers have certain demographics because we didn’t previously hire anyone different). If you can truly separate traits from demographics, you can make better decisions. Remember: They have to be objectively measurable.
RELATED: How to use survival analysis to predict when employees will leave
3. Act on the decisions
This is the part where HR can use the information and insights gained from predictive analytics to make the case for action. However, this is also the part where many HR managers get stuck. They have the data, they know what to do, but they find it difficult to execute. Think about it this way: Just as the Finance team can demonstrate why changing from one supplier to another will save the business significant money, you, as a representative of the HR team, can show how changes in the way their workforce is managed can affect the bottom line.
Using our attrition example, here’s how you might take action:
- Keep team sizes the way they are but implement some support for managers to make it easier to handle the administrative burden of larger teams.
- Review your onboarding process (using our onboarding checklist) and add additional support for new starters in their first year to set clear goals, build connections with their colleagues and identify high performers to fast-track development.
- Invest in more extensive training for high performers across the business helping them to broaden their skillset and use it in new ways which may help to reduce attrition and increase productivity.
- Realizing that satisfaction levels were lower in one part of the business, conduct an investigation into the root causes and solve them.
intelli tips: Making changes within your organization based on your predictive HR analytics data doesn’t need to be daunting. Here are a few things to remember:
- Start small – Prototype the solution, and build from there
- Communicate – We’re talking about basic change management here; make sure to bring everyone along for the ride.
- Be careful not to make more problems by solving the one you can see. At the same time, don’t just solve the symptoms; get to the core of the issue (use the data to check).
- Keep going! Track the changes you make and see the result. Every piece of data contributes to the next insight.
RELATED: Learn how Fujitsu General problem-solved their issues with employee engagement and goal setting and achieved an increase from 25% to 95% using intelliHR: Read the case study.
Predictive HR analytics helps you make informed business decisions – and fast
In conclusion, using predictive HR analytics can not only power good business decisions, but it can also give you the competitive edge to get ahead which is particularly important given that we’re working in a highly unpredictable time. It all comes down to setting up your processes to capture the data you need, having an objective, methodical approach to interpreting the results and then implementing the decisions effectively.
What next? Learn more about intelliHR’s advanced HR analytics tools and how they could help power your business decisions.