Using machine learning to keep your high performers

Machine learning is everywhere. It’s used in medicine, finance and e-commerce, and it’s become so sophisticated that it can predict both fraud and customer behavior. Fortunately for benefits leaders, predictive analytics can also help you further understand and care for your people. And according to recent research from Grant Thornton, this tool may be more valuable than ever.

A recent Grant Thornton State of Work in America survey shows that nearly one-third (29%) of employees are actively looking for a new job at a different company.

Using machine learning and your existing data will help you gain an even greater understanding of your people and their needs. You could even predict which employees are on the verge of leaving. And here’s the good news: Many companies already have the data they need to conduct these analyses.

Now, let’s discuss what to do with those numbers.

What data does machine learning need?

Companies already have easy access to employee salary history, performance ratings, and both disciplinary and commendation notes for every employee. Machine learning algorithms can also incorporate if an employee applied for an internal job opening, if they manage people, whether they have been flagged as “critical” talent, and if they have been identified as a high-potential employee.

Further, the machine learning required is not difficult compared with other artificial intelligence applications. Marketing professionals have used similar methods for years, with subscription-based businesses such as streaming services or mobile phone providers as two examples. These businesses embrace the concept that you can take customer data and deploy machine learning to prevent customers from leaving and persuade other customers to buy more.

In our work with Grant Thornton clients, we have seen machine learning predict which employees plan to leave a firm within six months or one year. For instance, while working with a large government agency, we used machine learning to accurately identify employees at risk of resigning or taking early retirement. This was done by reviewing several years of data provided by human resources. From there, we suggested proactive steps the client could take to extend these employees’ tenures and continue advancing their organization while keeping their people happy.

In other words, machine learning leads the way to win-win situations: success for your organization and happiness for your colleagues. Let’s explore what that looks like in practice.

Think like a marketer and use ‘Employee Preference Optimization’

Once a company identifies a high-performing employee who may be about to leave the organization, the focus shifts to deploying relevant retention tools.

Through tools like surveys, you can understand which benefits are most important to your people — and which you can enhance or add. This is called ‘Employee Preference Optimization.’ A well-designed survey will not just ask employees what they want directly, as employees typically say more of everything, but rather, ask employees to compare the perceived value in different packages — thereby understanding what they value most. In one instance, we found boosting the monthly car allowance for high-performing employees resulted in a perceived value almost double its actual cost.

More broadly, employee listening surveys allow for analyzing several benefits — ranging from health insurance coverage to retirement savings to vacation policy — in what is called conjoint analysis. The analysis is based on the perceived worth as a fraction or a multiple of the actual expense. Marketers use conjoint analysis to understand the relative value customers place on product or service features, and benefits leaders can use it to help measure an individual’s sensitivity to changes in benefits and rewards.

Related: Artificial Intelligence is the future (whether you like it or not!)

Employee surveys make it easier to perform effective analyses of the best benefits options for both employers and employees. Using these tools, an employer planning on boosting spending on rewards and benefits can be confident it is doing so effectively. It’s even possible to identify a total rewards package that 70% or 80% of employees believe is better than their previous package but costs your company thousands of dollars less per person per year.

It doesn’t get more “win-win” than that.

Tim Glowa is a principal and leader of employee listening and human capital services offerings at Grant Thornton LLP.