Precise attribute modeling uncovers 6.8% of incremental sales 

A global grocery retailer wanted its entire Merchandising function to be able to identify and understand the major drivers of financial performance.

Many decisions were based on individuals’ knowledge and experience, but without any consistent process, measures or accountability to understand when, and more importantly why, variance from the plan occurred.

We partnered with internal Analytics, Merchandising and Finance teams, to define and prove a hypothesis; that using available internal and external data, it was possible to accurately identify, isolate and quantify the key drivers of monthly store sales performance.

Our structured approach considered four key stages:

• Framing the challenge to clarify users’ key needs
• Creating the building blocks by creating a living database of over a thousand potential modeling attributes
• Developing a reusable and scalable capability by training our Performance Engine to model monthly category store sales and isolate the primary drivers 
• Validating the opportunity by demonstrating a highly predictive model which explained 86% of the sales variation, without using any historic sales input

Clear diagnostic scorecards prioritized the tactical actions for Merchandising and Stores to focus on, categorized by those which each party had control over, at the same time democratizing this insight to provide a single consistent view of the facts, from head office to stores.

Strategically, the ability to more accurately predict category performance provided insightful support to the annual budget setting process, ensuring that achievable targets were set across the business and identifying an incremental 6.8%pts of sales opportunity to achieve.

We uncovered 6.8% of incremental sales