A Case Study
Every day retailers collect and store volumes of transaction data totaling millions of transactions. Point of sale transaction data includes what items are purchased, by whom, how often, and from which source. There is more than enough information available to create a complete picture of how customers shop, but with the increasing use and acceptance of data storage, mobile devices and online shopping, the flow of data is so great that most retailers don’t have the technological and human resources to organize and leverage that data.
Obstacles to leveraging data:
Speed and volume of transaction data
Analysts and data scientists to interpret the data
Ability to convert information into sales
Given their resources, most retail organizations have exercised very little of the data they create and collect. Historically, retailers have been able to use some data to create product bundles or make adjustments to display strategies. Dashboard and BI technologies have allowed some retailers to ask questions about their data, giving them a limited glimpse into sales patterns.
Retailers need to:
Create display and layout strategies that drive sales
Optimize inventory based on buying patterns across product categories
Reduce lost sales due to lack of availability of complementary products
Detect product substitution patterns from buying behavior
Leverage product substitutes to reduce inventory while maintaining service levels
EmcienPatterns software automatically uses point-of-sale data to create a deeper understanding of shopping patterns and trends. Retailers will see what products are redundant, which products can function as substitutes, which products are purchased together or in bundles, and which products simply aren’t selling.
Automated Analysis helps retailers:
Increase transaction size
Eliminate lost sales due to lack of availability
Increase the purchase of complementary products
Reduce inventory through SKU reduction
Understand regional sales trends
The entire category management process has been driven based on a siloed mentality and a data management approach that lacks a holistic view of the product offerings and without meaningful, detailed insight into customer buying patterns and behavior. Customers do not buy exclusively within the category silos. However, due to the limitations of data processing, the category level has been the focal point of all retail analysis to date.
Emcien’s algorithms compute all the possible combinations that customers have bought within and across categories then rank the item groups based on popularity and frequency of purchase. This presents a ranked view of the most popular item groups across the store. Emcien then uses a graph data model approach, revealing the relationships between items across the store. This holistic approach allows you to understand how the items in your product mix connect, and how customers are behaving in your store based on your mix.
The output from Emcien’s pattern detection engine can be used to drive downstream planning activities or as an input to enhance downstream systems such as:
- Product mix planning – Inter and intra category dependencies.
- Product planning – Product kits by location.
- Inventory planning using product substitutes.
Financial and Operational Impact
Typical clients in retail have seen the following financial and operational results:
- 2% increase in total sales due to improved product mix planning by including impact of cross-category interactions.
- 12% inventory reduction with no lost sales.
Emcien’s pattern detection platform connects the dots between all products, customer segments and transactions, giving retailers, distributors and consumer product companies actionable insight into each and every item in inventory based on buying patterns.