Forecasting Future Personnel Requirements Using Machine Learning
A practical demonstration of using machine learning to predict personnel requirements based on customer volumes
By | Adam D McKinnon | www.adam-d-mckinnon.com
The goal is to know what’s coming…
Around this time last year I began immersing myself in Forecasting. My learning and professional experience since suggests to me that there are considerable opportunities to apply Forecasting methods in people related decision-making. I see Forecasting informing decisions such as: – retail rostering through the forecasting sales volumes, – workloads for talent acquisition professionals through forecasting requisitions volumes, – job design by forecasting volumes of service delivery staff, – Strategic Workforce Planning (SWP) through anticipation of skill demand, and – many more (I suspect imagination may be the limiting factor)!
While the last article I wrote on Forecasting was fun, this article adopts a more pragmatic stance by providing a practical demonstration of forecasting customer contact volumes to manage staff rosters.
Business Context
The Brisbane City Council (Australia) resolves hundreds of customer contact enquiries by residents each day. In a bid to ensure ongoing customer centricity, the council offers multiple channels through which residents can contact the council. It is expected that council staff will resolve a particular number of customer contact enquiries each day. The rate of resolution varies by channel and is measured as follows (these rates are fictitious):
- Email: 175 / employee / day,
- Face to Face: 75 / employee / day,
- Mail: 75 / employee / day,
