Node risk model

Sterling Intelligent Promising Premium uses a node risk model to predict the probability and cost that is associated with risk of order cancellation and reschedule. As a fulfillment manager, you can use the node risk model to enhance the efficiency of inventory management.

Various factors can affect the fulfillment of online orders. Order cancellation and reschedule from the seller or shopper are two of the common issues that the fulfillment managers face that result in inefficient order fulfillment. Orders get canceled or rescheduled due to many factors:

Customer initiated order cancellation
Inefficient inventory management
Low node capacity
Backlog at the node
Shipping difficulties for heavy products
Misplaced products

When the order fulfillment is impacted by these factors, a fulfillment manager can use the node risk model to predict probability of order cancellation and reschedule. The node risk model analyzes historical data to predict the probability of order cancellation and reschedule. The model then converts the predicted cancellation rates and reschedule rates into costs.

The node risk model predicts order cancellation probability and order reschedule rates by using artificial intelligence (AI) to process the historical data from various data feeds.

Note: To enable the node risk model, you must upload data for the following data types:

Order line data
T-log data
Markdown data
Availability data
Catalog item data

For more information about transferring the data for the mentioned feed types, see Data feed types overview.

The node risk model uses AI to predict cancellation and reschedule rates and calculate the respective costs.

Predicting cancellation probability and cost: The model uses AI to predict the possibilities of order cancellation for a node. A fulfillment manager can get insights into the order cancellation possibilities at a specific node and the historical reasons of order cancellation.
Either the seller or shopper can cancel the order for various reasons. Every time an order or an item from the order line is cancelled, the cost of associated with the loss of sale is considered as cancellation cost. The node risk model converts the predicted cancellation probability into cancellation cost.
Predicting reschedule probability and cost: The model uses AI to predict the probability of order reschedule for a node. A fulfillment manager can get insights into the order reschedule possibilities at a specific node and the historical reasons of order reschedule.
Whenever a node cannot fulfill an order. The fulfillment manager can reassign the order to be fulfilled from another node, which results in rescheduling of order fulfillment. Every time an order or an item from the order line is rescheduled, multiple costs that are associated with it, such as delay cost, node-balancing cost, cost associated with change of shipping method, are considered as rescheduling cost. The node risk model converts the predicted reschedule probability into reschedule cost.

Based on insights of rates and costs of cancellation and reschedule, the fulfillment manager can adjust the optimization profile. For more information about creating or adjusting the optimization profile for the expected node performance, see Managing optimization profiles.