Coping During High Demand Season
Managing production during the peak demand season has been a challenge for Skretting more demanding than usual. Unforeseen short-term changes, pressure from major clients, a highly volatile supply-chain, and experienced people from the planning department leaving the company all added to the issues. Furthermore, given the 2-week visibility on demand provided to them through the efforts of the sales team, as well as unforeseen production interruptions due to high-priority orders, Skretting was operating solely on a Make-to-order (MTO) strategy, shifting production from day-to-day to accommodate for the incoming changes.
Need of the hour
As such, the company was in need of a solution that would
- allow them to extend their ability to anticipate demand beyond the current horizon,
- harmonize Make-to-stock (MTS) and MTO, making use of surplus production capacity during low-demand season to improve production planning and relieve pressure during the high-demand season.
A Two-Phase Approach
In order to improve this situation, a two-phase plan was outlined:
- The initial phase focuses on building an interactive MTS recommendation engine, allowing production planners to get insights regarding the demand in the upcoming period as well as optimizing production runs.
- The second phase takes a deeper look into Skretting’s existing forecasting in order to develop a more robust demand forecasting platform that will in turn reinforce the recommendation resulting from the 1st phase.
As the initial phase began rolling out, it quickly became clear that a crucial first step in designing a successful system to provide actionable MTS-MTO recommendations is to select a suitable subset of Stock keeping units (SKUs) to be considered in the underlying decision-making process. Through regular interactions with the Skretting team, a large number of SKUs were excluded a priori from the scope of this project, as their nature makes them unsuitable for an MTS production strategy.
Furthermore, analyzing a variety of historical data including sales, manufacturing, inventory and margins, allowed the team to engineer several relevant features that would form the basis of the MTS selection process. After clustering and classifying these items and identifying product groups with favorable MTS characteristics, the recommendation engine was approached as an optimization problem, suggesting an optimal production volume given the existing production and capacity constraints.
As for the improved forecast methodology, a number of statistical and machine learning models were considered and trained on the available historical sales. More specifically, different variations of the Seasonal AutoRegressive Integrated Moving Average (SARIMA) model, which attempts to predict future sales based on their history, were taken into account with employing a champion approach and selecting the best performing model on an Item-by-Item basis.
The final solution presented to Skretting consists of an interactive MTS Pro application, allowing the firm’s production planners to visualize and inspect the recommendations coming out of the MTS engine as well and offers the possibility to explore the improved underlying forecast.