Delivering an effective cost prediction model and actionable analytics
ASR Group built a POC that includes a predictive algorithm to forecast trucking costs more than 30 days out. The model uses 12 months of internal company data and a descriptive algorithm based on seven sources of external data. These include economic data, labor statistics, transportation statistics, weather information, transportation management system data, and more.
SAP Data Intelligence Cloud integrates, orchestrates, and manages the data and the predictive algorithm, which provides a network-level expression of costs. The confidence interval is based on:
• 12 months of data (maximum)
• Lane network ratio
• Lane level prediction
• Three-month moving average for freight price data
• First derivative calculation of freight price by lane
• Measure of the distance between outputs from the long-term forecast to the near-term moving average
• Conditional logic that selects the best output
SAP Analytics Cloud is then used by the business for visualization and what-if analysis.