NRM and Promotional Insights Automation

Synopsis

Trusted Data were asked to support a global FMCG player in the UK, who had encountered constraints in mapping promotional effectiveness across its’ mainstream retail channel. Investment in a generalised Trade Promotion Management solution from a global consulting firm had failed to provide requisite / reliable insights and manual analysis conducted in excel was too cumbersome and lacked granularity.

Baseline sales computation in a heavily promoted category was not possible through more conventional analysis and inferring competitor promotions and related cannibalisation metrics had not been achieved, as in-market reports on promotional activity captured less than 10% of category promotions.

Blending multiple data sources to create a singular source of truth for the Client was earmarked as a crucial milestone and given the variety in promotional processes at the account level, a software solution could not offer the Client dynamic and reliable insights.

Client Overview

🔸Supplying all major UK grocery retailers in a heavily promoted category 

🔸Current TPM solution was redundant and not being used by commercial or analytics team

🔸Leading Global Food and Beverages brand

Client Problem

Transforming retail promotional performance from struggle to success; our client had struggled with manual, cumbersome and somewhat inaccurate analysis, which had hindered their understanding of promotional effectiveness. They then embraced an AI-based approach that revolutionised insights automation. With enriched data and insights, greater accuracy, and minimal manual intervention—all at the click of a button—our team enabled the client to acquire unprecedented insights into category performance, promotional impact and competitors“.

 

Analytics Delivery

Modelling required a robust data pipeline being created to call in data from multiple sources and systems in various formats, including promotional calendar information, product-level financial metrics and Nielsen sales data. Following extensive data management and transformation work, we then set about to:

🔸  Autonomously compute a dynamic baseline by SKU for any time period

🔸  Dynamically infer all deals in the category, including competitors

🔸  Provide rapid insights on Promotional and Sales KPIs

 

The Challenge

⦿   Heavily promoted category and unable to reliably compute baseline sales demand for the clients’ products

⦿   Unable to detect competitor promotions from available data in a reliable way

⦿   This had a knock on effect as linked KPIs such as promotional incrementality, cannibalisation and switching could not be calculated

⦿   Data was disparate and uncertainty cross-functionally on the correct data to use for analytics

⦿   Each SKU had a different product ID between datasets, which inhibited the clients’ ability to create linkage between datasets for SKU-level NRM analysis

Trusted Data Approach

⦿   3 – 4 weeks to map internal processes, expectations and solution needs

⦿   Amalgamating disparate data sources to centralise promotional data in a robust pipeline

⦿   Mapping each product to a central tag, to cross reference between datasets

⦿   Machine learning approach to compute a dynamic baseline for every product, read all category promotions and dynamically compute cannibalisation versus similar competitor products for any promo

⦿   Automated code to compute Promotional indices and KPIs, including Promo sales uplift, Cannibalisation and Promo ROI

OUR IMPACT

Click Promotional Insights

%

Reduction in Manual analysis

%

Reduction in leadtime to gather Quarterly Insights