Competition for skilled tech workers is fierce, so a new program actually predicts when an employee is considering resignation, and how you can implement retention.
Crystal balls, fortune cookies and tellers, and, sadly, our own perceptions, can’t predict the future. This indecision may lead your coveted employees to ask themselves the question The Clash posed in 1981: “Should I stay or should I go?” Your company’s go-to girl(s) or guy(s) are a hot commodity, and it’s hard for them to ignore how they and their tech skills are in heavy demand. Can the lure of a competitor’s beckoning be stilled? Perhaps.
Combatting the competition
Unemployment is low, 3.6% as of October 2019 (up from a record low, since 1969, of 3.5% in August 2019).that many open jobs remain unfilled for weeks, even months. So when a company can boast a happy, skilled tech worker, the company can also claim success. Unfortunately, you can’t rest too long on your laurels.
Seeing into the future
Employers can offer competitive salaries, great benefits, and work to , but it’s much more difficult to know how someone is feeling, let alone predict what their future actions might be. You want them to stay. They might think about leaving. It’s a common concern shared by many tech departments and companies. With this in mind, UK-based KPMG developed Workplace Analytics to implement retention solutions as soon as an employee is identified as considering leaving. Most solutions identify systematic problems that drive attrition, but KPMG says their solution combines internal and external data to provide actionable insights to retain individual employees.
SEE: Hiring kit: Chief diversity officer (TechRepublic Premium)
The proprietary, managed Workplace Analytics evaluates hundreds of behaviors such as email use, and phone, travel habits, commute times, and paid time-off (PTO), while considering unemployment and opportunity in the surrounding area, to provide individuals’ retention scores. As that score fluctuates, an employer can take appropriate actions to counter the likelihood that an individual will quit. For example, the employer can promote workplace flexibility, adjust workloads to avoid burnout, or even offer perks.
Noticing subtle behavioral changes
“No matter the industry,” said Bill Nowacki, managing director of Decision Science at KPMG, “each employee within an organization has various normative behaviors, and the solution is designed to spot slight deviations from that behavior. Ultimately, these deviations are typically indications of a ‘restlessness’ that precede job change.”
As an example, Nowacki explained, “I arrive to the office right around 6 am every weekday and you’ll see me on my VPN every Saturday and Sunday morning, but let’s say I have a couple of new job opportunities in the pipeline and I’m far along in the interview process for one, so I stop showing up at 6 am and no longer work weekends, this you could say would be a deviation from my normative behaviors, and these deviations typically preface that something is going to happen.”
That enterprise-wide learning curve
Having predictive software is essential in the tech industry, he said. “Because the technology industry is often subject to high turnover due to strong competition, specifically in a time when unemployment is at all-time low, there’s a lot to learn specifically from big tech. While the industry doesn’t benefit any more or less than other industries, one difference is that employees within the technology industry tend to be more fluid. This solution gives organizations the ability to spot an employee’s change of heart early on and empowers them to act quickly to solve the issue both at the individual and enterprise-wide levels.
“In other words, if ain the engineering department resigns you can start to dig into the ‘why.’ Eventually the ‘why’ she left becomes clear when you find that your company pays 5% less than the general market. Using these insights, you start digging a bit deeper and see that many python programmers throughout the organization are starting to show signs of dissatisfaction. This allows us to begin understanding why whole groups of people appear to be turning over at a rapid pace and enables us to identify the systemic problems that are fueling those departures and preventatively resolve them.”
Beware of ambivalence and the disconnect
The signs, he furthered explained, are noticing someone disconnecting or becoming ambivalent about work. An organization can also check on an employee’s length of day, if it’s “static or shortening,” check how many badge swipes they’ve made, their Diner Club card-usage, first email of the day, number of phone calls made. “By analyzing the furthest two points we’d then be able to determine the length of day. While it’s true that we might be able to get a sense of length of day through one of the sources, it’s crucial to stitch together all of this big data fabric to get a complete and accurate picture and that’s what we’re doing with Workplace Analytics.”
The top 5 analytics to leverage
Nowacki said they have identified 1,500 signals helpful to identifying employees in this equation, and added that the company’s program has “five particular analytics we leverage so that organizations can take action to resolve the issue:”
- Identifying attrition and its reason before it happens.
- Identifying burnout as it’s happening.
- Identifying disconnect as it’s happening.
- In an incoming cohort or class, figuring out who the rock stars are, so an organization can accelerate their learning and nurture them.
- Discovering and understanding why the perennial top performers are outperforming the others in a cohort, so an organization can better coach the broader group.
Nowacki said they originally used “survival” analysis to solve problems as it is transparent and identifies things that increase the hazard score (the hazard is the employee quitting), which can be 70 to 80 reasons why a person wants to leave. At this point, the company has HR meet with the employee to analyze the reasons they want to leave and try to come up with solutions. He added, “Of course, the remediation can’t always be money, but there are other solutions like a more tailored benefits-package or childcare at work, that might do the trick.”
The programs’ origins, he explained were developed to assess the solution to target student retention in UK universities. It’s now been applied it to HR departments. “The Workplace Analytics solution works best in industries that have discrete measures of productivity and engagement,” Nowicki said. “For example, manufacturing has a lot of measurements—such as calculating cost of labor or days lost to injury on a factory floor—which can be used in various machine learning models. Whereas, in contrast, an accounting firm might find it harder to measure the true cost of employee disengagement.”
“Similar to the idea that some industries have a higher volume of measurement inputs than others, there are certain roles within a company that leave less “digital exhaust,” as we like to call it, than others. Imagine a cashier at your department store, she only has a handful of parameters that led us to understand her baseline productivity, such as clock-in and clock-out time, so it’s harder to identify when she starts to stray from the norm. In general, some industries lend themselves better to this solution because job roles and positions inherent to said industry are better measurement vessels.”
Monthly analytic aid
And because it is a managed monthly subscription system, designed for companies of all sizes, there is no learning curve for the company it’s serving. “We score each employee in a customer company on a regular basis based on the five criteria, attrition, burnout, disconnect, etc., and then send results and proposed solutions to supervisors,” he said. “This means that key stakeholders in this situation, mostly supervisors with direct reports, only need a modest amount of training on how to handle alerts and to take the action recommended in the workflow report.”
For more, check out How to conduct effective one-on-one meetings on TechRepublic.