Workforce optimisation for DFS
We used machine learning to predict hourly footfall in store, and then optimisation to produce workforce schedules that better met supply with demand, improving conversion rates and employee fairness.
Challenge
As market leaders in the home furnishing industry, DFS continues to invest heavily in both operational, and customer-facing technologies. To ensure the best possible in-store experience, and to better meet customer demand, DFS chose Satalia to build their workforce optimisation system.
Outcomes
18%
increase in conversion rate
£230K
additional revenue (1 store, over 11 week period)
77%
hourly prediction accuracy, 10 weeks in advance
1 hour
to produce 10 weeks of predictions across 117 stores
Outcomes
18%
increase in conversion rate
£230K
additional revenue (1 store, over 11 week period)
77%
hourly prediction accuracy, 10 weeks in advance
1 hour
to produce 10 weeks of predictions across 117 stores
Solution
The system consisted of two components. First, a demand forecasting model which used historical footfall, sales data, seasonality and weather data to produce hourly forecasts for all 117 of their stores, 6 weeks in advance. The second was an optimisation model, which used these forecasts to produce optimised staff schedules, suggesting the exact employees needed to serve hourly demand.
A key metric for DFS was fairness. Sales representatives earn commission for every sale, so those who are consistently allocated to busy shifts have more earning potential than those who aren’t. Using our forecasts, we produced schedules that ensured everyone had the opportunity to work the busier shifts. The forecasts are also being used to review (and potentially adjust) their shift patterns and opening times, with the plan of building a more contingent, flexible workforce alongside their existing employees.