Optimization of markdown schedules for clearance items at physical retail stores

Patent metadata

Field	Value
Publication number	US-2021304243-A1
Application number	US-202016835917-A
Country	US
Kind code	A1
Filing date	Mar 31, 2020
Priority date	Mar 31, 2020
Publication date	Sep 30, 2021
Grant date	—

Abstract

Methods and systems are described for optimizing markdown schedules for clearance items at physical retail stores. For example, price sensitivity of a current inventory item at a physical retail store may be modeled using techniques that account for differences in operating behavior and data availability at physical stores versus online channels for selling items. When the current item is placed on clearance at the physical retail store, a request for a markdown schedule, including goals for the clearance, may be received. The forecasted price sensitivity of the current item may be adjusted based on actual price sensitivity of one or more past clearance items (e.g., based on actual clearance sales data) determined to match the current item. An optimal markdown schedule for the item may then be determined based, at least in part, on the adjusted price sensitivity of the item and clearance goals.

First claim

1 . A method for optimizing an inventory item markdown schedule, the method comprising:
building a first model that represents a forecasted price sensitivity of a current inventory item at a physical store of a retailer; receiving a request for a markdown schedule when the current inventory item is to be placed on clearance at the physical store; generating a set of potential markdown schedules based on parameters of the request, each potential markdown schedule of the set defining one or more discounts and an associated discount duration; identifying a past clearance item that matches the current inventory item; applying a second model that represents an actual price sensitivity of the past clearance item to the first model to adjust the forecasted price sensitivity of the current inventory item; and determining an optimal markdown schedule from among the set of potential markdown schedules based, at least in part, on the adjusted price sensitivity of the current inventory item.

2 . The method of claim 1 , wherein building the first model comprises:
enhancing price data associated with the current inventory item at the physical store; determining a demand for the current inventory item at the physical store based on the enhanced price data; clustering the physical store with one or more other physical stores having stock-to-sale ratios comparable to a stock-to-sale ratio of the physical store to form a current item-cluster, wherein price and demand data determined for the current inventory item at each physical store within the current item-cluster is aggregated; and based, at least in part, on the aggregated price and demand data, estimating a price elasticity of demand of the current inventory item for the current item-cluster, wherein the forecasted price sensitivity of the current inventory item is measured using the price elasticity of demand.

3 . The method of claim 2 , wherein determining the optimal markdown schedule further comprises:
determining a number of units of the current inventory item for allocation to the physical store based on one or more of current inventory levels and pre-clearance demand for the current inventory item across each of the physical stores in the current item-cluster.

4 . The method of claim 2 , wherein enhancing the price data associated with the current inventory item at the physical store comprises:
aggregating price data for the current inventory item at the physical store with price data for the current inventory item at one or more other physical stores similar to the physical store.

5 . The method of claim 2 , wherein determining the demand for the current inventory item at the physical store comprises:
determining an effect of a stockout of the current inventory item on the demand; and adjusting the demand based on the determined stockout effect.

6 . The method of claim 5 , wherein determining the stockout effect comprises:
based on historical clearance data, determining a subset of physical stores from a totality of physical stores of the retailer having a minimum threshold of sales; for each physical store in the subset of physical stores, determining a percentage change of sales at one or more clearance discount percentages; for each unique clearance discount, generating a distribution, wherein the distribution includes a minimum and maximum value representing a minimum and maximum percentage change allowed at the respective clearance discount.

7 . The method of claim 6 , wherein if the physical store for which the request for the markdown schedule for the current inventory item is received is constrained by inventory at one or more of the unique clearance discounts, the method further comprises:
randomly generating a first value between the minimum and maximum value of the distribution for the one or more of the unique clearance discounts; randomly generating a second value; adjusting a number of sales of the current inventory item at the physical store based on the first randomly generated value when the second randomly generated value is greater than a predefined value; and determining the demand for the current inventory item based on the adjusted sales.

8 . The method of claim 2 , wherein identifying the past clearance item that matches the current inventory item comprises:
identifying a past item-cluster comprised of a cluster of physical stores at which the past clearance item was sold that matches the current item-cluster based on one or more of a time of year, a stock-to sales ratio, level of inventory, pre-clearance sales, a number of stores, and a clearance program length.

9 . The method of claim 8 , wherein applying the second model that represents the actual price sensitivity of the past clearance item to the first model comprises:
retrieving, as the second model, a model built based on data from the past item-cluster including actual clearance sales data for the past clearance item at the cluster of physical stores.

10 . The method of claim 1 , wherein more than one past clearance item is identified as matching the current inventory item, and adjusting the forecasted price sensitivity of the current inventory item comprises:
for each of the past clearance items, applying a model that represents an actual price sensitivity of the respective past clearance item to the first model to adjust the forecasted price sensitivity of the current inventory item.

11 . The method of claim 1 , further comprising:
determining the optimal markdown schedule from among the set of potential markdown schedules further based on one or more clearance goals included within the parameters of the request.

12 . The method of claim 1 , further comprising:
updating the optimal markdown schedule for the current inventory item at the physical store at predetermined intervals throughout a clearance period for the current inventory item.

13 . The method of claim 1 , further comprising determining a unique optimal markdown schedule for the current inventory item for each of a plurality of stores of a retail enterprise.

14 . A system for optimizing an inventory item markdown schedule, the system comprising:
a processor; a memory communicatively coupled to the processor, the memory storing instructions that, when executed by the processor, cause the system to:
build a first model that represents a forecasted price sensitivity of a current inventory item at a physical store of a retailer;
receive a request for a markdown schedule when the current inventory item is to be placed on clearance at the physical store;
generating a set of potential markdown schedules based on parameters of the request, each potential markdown schedule of the set defining one or more discounts and an associated discount duration;
identify a past clearance item that matches the current inventory item;
apply a second model that represents an actual price sensitivity of the past clearance item to the first model to adjust the forecasted price sensitivity of the current inventory item; and
determine an optimal markdown schedule from among the set of potential markdown schedules based on the adjusted price sensitivity of the current inventory item.

15 . The system of claim 14 , wherein the instructions cause the system to determine the optimal markdown schedule from among the set of potential markdown schedules using one of a grid search or a reinforcement learning technique.

16 . The system of claim 14 , wherein the first model is built using a plurality of data inputs, the plurality of data inputs including one or more of price data

Assignees

• Target Brands Inc

Inventors

• Le Luyen
• Bhattacharya Saibal

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06N20/00

  Machine learning · CPC title

• G06Q30/0235 Primary

  constrained by time limit or expiration date · CPC title

• G06Q30/0223

  based on inventory · CPC title

• G06Q10/067 Primary

  Enterprise or organisation modelling · CPC title