How exactly does Data Analytics help Solve Worker Turnover?

The phrase “data analytics” — along with ‘digital transformation’ and ‘AI solutions’ — appears everywhere these days. It promises a better tomorrow with higher profits and happier people. Organizations in sectors from healthcare to publishing to farming to finance seem on a race to build a team of data scientists. They hope that the mere existence of this team will automatically translate to better business outcomes.

However, according to a recent analysis published by MIT Sloan Management Review, “Data analysts often fail to produce insights for making effective business decisions [because] Leaders need to make sure that data analytics is decision-driven.”

In other words, instead of broad data analytics endeavors where you hope to find patterns and purpose from your data, focus on the decisions you need to make and find data for that purpose.

Decision-Driven Data Analytics

Workforce decisions are among the most impactful decisions an organization makes. The matching of a person to a job — whether a new hire or an internal re-assignment or an employee promotion — has the potential to strengthen staff morale and drive business outcomes more effectively. It also has the potential to accelerate turnover, spiral staff burnout, and produce poor business outcomes.

The solution? Optimize for the outcomes that matter. Gather relevant outcome data, align it to people across the organization, and use that to inform workforce decisions. Sounds simple? Yes. Easy to do? Not so much.

With 10 years experience collecting outcome data and leveraging machine learning techniques to match people to positions where they will thrive, Arena Analytics has some insights to get you started. To build an analytics backbone to strengthen your workforce, begin with these reports:

  1. Milestone turnover
  2. Remaining tenure inflection points
  3. Turnover trends by prediction scores
1) Milestone Turnover

Let’s start with an actionable definition of turnover. The simple employee turnover metric we see every day is far too broad to yield any clear follow-up actions. 

When you unpack turnover data, however, conversation shifts and targeted solutions arise.

Take this one view of turnover data.  Of the total number of people who left in a given time period, what percent were employed less than 3 months, 3-6 months, 2-5 years?  Now imagine this same pie chart broken down by job function, by department, by location.  If the largest percentage of turnover occurs within 3 months of being hired, the problem may lie in the initial hiring decisions.  

Re-hiring repeatedly for the same role, several times over the course of the year costs far more than re-hiring every 5 years, or even every 2 years. A focus on improving the short tenure rates (those in red above) may yield the largest savings and have the greatest impact on staff morale.

We call this defined tenure timeframe a Milestone-Adjusted Turnover Rate. The image below shows how you calculate a 3-Month (or 90-Day) milestone turnover rate by following a cohort of hires to see how many leave within 3 months of their hire date.

2) ‘Remaining Tenure’ Inflection Points

Another view of turnover looks at retention. This chart below shows the Median Remaining Tenure for same position, at different locations at a healthcare system.  The x-axis tracks each day a person works the job, while the y-axis shows the average (median) remaining days this employee stays on the job.  For the red line, the downward trend reveals a steady decrease in tenure. As each day passes, there are fewer days remaining. The blue line shows the opposite, the longer someone stays, the longer they will continue to stay.

Collaborative discussion can explore reasons behind these differences and ways to improve the red line location’s retention. Follow up meetings can continue to examine how these trends shift in response to new policies, procedures, and people.  Below is an example for a different position and place of work where retention does not trend up or down. Instead, there are strong inflection points. Armed with this insight, employers can step in at Day 100 and then again Day 200 to address pending dissatisfaction before it fully develops.

3) Turnover Trends by Prediction Scores

Arena Analytics’ machine learning platform encapsulates the analytics above and many more data points to create data pictures of each individual.  With 10+ years of outcome data and ongoing data feeds from each client, Arena can predict whether or not new candidates (or existing staff) will fall into the red or green lines below for each open job. 

Operational and HR leaders can then look at the prediction scores and move away from gut checks and theoretically ‘good’ behavioral traits.  By closing the gap between HR and Ops with data, organizations can focus on achieving their greater mission – quality care, excellent service, and optimal revenue.