Travis Perkins is a leading British builders merchant and home improvement retailer operating 21 distinct businesses across its branches, with more than 100,000 stock keeping units (SKUs) to manage.
Travis Perkins wanted to reduce inventory while maximizing availability. New capabilities were needed to optimize store inventories and analyze product performance across many attributes through analysis of huge volumes of data. More than 100K products, substitute products, and 1000+ branches of multiple types made this a daunting problem.
Core Compete developed a comprehensive bespoke solution for weekly performance reviews and analysis across many attributes. The new solution enables end-to-end annual line reviews, collaboration, and ‘what if’ analysis to simulate product change impacts throughout the supply chain, store-SKU assignments and direct integration into replenishment systems.
To drive maximum value from the big data and analytics solution, Core Compete brought together the right components in four critical areas.
Domain Expertise – Our retail solution architects designed a data lake that integrated large volumes of data including Invoices, Daily Inventory, Customer Master, Planograms, and Trade Promotions. A comprehensive data model enables dynamic insights spanning sales, market basket, and performance related to inventory, promotions, planograms, and branch-specific metrics.
Analytics Innovation – Several retail analytics areas were integrated to facilitate end-to-end assortment workflows including Assortment (product-branch) scoring, market basket analyses, and automated outlier detection and elimination to compute standard rates of sale for products.
Technology Know-How – The solution was deployed on a Hadoop Data Lake on AWS. Spark was used for automated aggregation routines to address complex large-scale calculations. SAS platform enables visualizations and analytics. Core Compete provides 24/7 management of the environment allowing the client to focus on their business.
Agile Delivery Methodology – The solution was delivered in phases over a 9-month development and delivery cycle. Within 12 weeks, the data lake was deployed with more than a dozen weekly performance and attribute-based analytics dashboards. Subsequent phases addressed planogram management and comprehensive annual line review workflows.
Travis Perkins has moved closer to the goal of reduced non-productive inventory through a deeper understanding of range performance that drives effective actions. Another big benefit is the elimination of multiple weeks of latency through instantaneous impact analysis of range changes on the supply chain. Changes to the range can now be executed far more easily with more collaboration.