How AI Will Fight Turnover — And What You Can Do Until You Have It

Press enter to search
Close search
Open Menu

How AI Will Fight Turnover — And What You Can Do Until You Have It

By Ashish Gambhir, Cofounder and President at ShiftOne - 08/13/2019

From Deep Blue, the pioneering IBM chess program that defeated world champion Garry Kasparov in 1997, to contemporary machine learning algorithms that construct faux human faces or identify wind energy patterns, the power of artificial intelligence (AI) has captured society’s attention — and wracked its nerves. 

Yet everything to date is a surface scratch. With machine learning tools becoming more ubiquitous by the day, their applications have drifted into a more practical domain.

A prime example: what if we could leverage the superhuman processing abilities intrinsic to machine learning to tackle our industry’s foremost business dilemmas? It’s hardly a groundbreaking thought: in the corporate sphere, machine algorithms are already easing the customer support desk process and enabling faster contract reference, among other things.

But that’s mechanical stuff. What about the human problems? What about employee turnover, the scourge of the restaurant industry?

The turnover rate among hourly restaurant workers in America is well over 100% per annum, and the cost of an individual team member’s departure is anywhere from $2K to $5.6K when training, opportunity cost, etc. are factored in. For enterprise brands with thousands of hourly employees, that figure quickly scales into the millions and beyond.

This colossal inefficiency costs businesses billions of dollars in revenue every single year. Brands have thrown everything they can at it: generous vacation policies, beefed-up compensation structures, health insurance, loan programs, parental leave, etc. This culture and benefits double down works for some organizations: Chick-fil-A, one of the nation’s most respected fast-food brands, boasts a turnover rate 40% lower than the national average. 

For most, though, turnover remains a massive — and expensive — quandary. AI could be the answer.

It all starts with data. Each employee generates a tremendous amount of information: sales data, guest satisfaction surveys, scheduling data. And as with any data set, there are trends to be uncovered.

When we overlay team member data with turnover data and pour everything into a machine learning program, it becomes clear that strong links exist between the two.

Even without programmed AI tools, however, operators can follow simple steps to keep their top performers.

 

  1. Stay apprised.  Data is powerful, but in order to extract any kind of insight an operator must be both literate in the numbers and constantly tapped into the pulse. Pull regular reports on every metric that’s important to your business — sales, guest satisfaction, even prep or drive thru time — and track performance at the restaurant and individual level. 

  2. Watch for trends.  Although it’s easier for a computer to detect than a human eye, all data contains trends. In time, an operator keyed into crucial performance metrics will notice telltale signs of employee dissatisfaction — sales dips, neutral or negative reviews, decline in shifts. 

  3. Be proactive. When these trends are observed, don’t sit back and wait to see what happens. If an operator suspects an employee — particularly a top performer — is burning out, he or she must make the opening move. It’s as simple as a single question: “How are things?” or “Are you happy here?” A five-minute conversation could save you a rockstar — and $5K in hiring and training costs.

 

About the Author

Ashish Gambhir is a 15-year veteran of the employee engagement sphere. He is the president and co-founder of ShiftOne, formerly MomentSnap, a mobile app designed for hourly workers that reduces turnover, drives incremental revenue, and makes work a better place for hourly employees and managers. It uses real-time, individual performance data — usually from a point-of-sale system or guest review collector —  to power competitions, encourage communication and recognition, and monitor operational health factors like attrition risk.