In an intensely competitive retail environment of blurring channels, and ever-evolving consumer purchasing behavior, retailers are finding it more difficult to align their assortments to customer demand; a challenge that is magnified for merchandise offered in size ranges. Therefore, achieving optimal recommendations on location-specific size range profiles to better inform and align buying and inventory allocation decisions to consumer demand is critical.
Customers come in all shapes and sizes and serving their fit preferences, efficiently and effectively, at their desired store or channel is a key driver of profitable retailing. In order to satisfy each of these customers, location-specific size profiles play a key role throughout the buying and inventory management processes. Size profiles should influence buying and initial allocation decisions during the assortment planning or buying processes, yet buyers and planners often only manage buying decisions at the style-color/store cluster levels and expect that analytics will intelligently explode that plan down to size/dimension (SKU) and store. This explosion is necessary to execute the assortment plan through SKU creation, purchase orders, and initial allocation.
The ultimate goal in fashion retailing is to create store-specific assortments to best align inventory and buying decisions with customer demand, thus driving a greater return on inventory investment. Understanding each store’s unique size selling patterns and therefore size need by merchandise area and attribute is a critical step to meeting that unique customer demand. The massive amounts of data and labor-intensive tasks to analyze the data often paralyze the buyer and/or planner and drive them to settle on department and chain level averages.
Retailers have continued to look back at history in order to make future planning decisions. But historical size selling is often a result of misplaced inventory that results in lost sales and excessive markdowns. Retailers need a “cleansed” history that understands lost sales, exceptional sales, and the size demand potential. Without this “cleansed” historical performance, buying and allocation decisions will be based on last year’s mistakes, thus repeatedly disappointing customers and leaving unwanted sizes in some stores or selling channels.
Oracle Retail Profile Science provides optimal recommendations on location specific size range profiles to better inform and align buying and inventory allocation decisions to consumer demand. Oracle Retail evaluated its next-generation size profile solution at a fashion retailer and found these results.
Oracle Retail Profile Science identifies location-level selling patterns across different size ranges to systematically create accurate profiles of size distribution by merchandise category by location. It corrects for out of stocks and lost sales, uses robust simulation techniques, and identifies the right levels to generate the profiles. In addition to systematically generating the profiles, the solution allows for profile management, including comparison and analysis views, with an exception based management and approval workflow.
A new DataScience.com article gives additional examples of how Oracle Retail Profile Science, available as part of the Oracle Retail Science Platform, makes use of some classic machine learning techniques to estimate size profiles from a retailer’s historical sales data.