Ganit's Recommendation Engine helps one of
the largest CPG companies to identify additional
revenue channels with its retail partners

INDUSTRY OVERVIEW

The monthly demand at the retail end point is influenced by three factors:

• Retailer's Purchase Potential, reflective of Historic Demands in every region.
• Changes in Consumer Behaviour.
• Conscious effort from the retailer to improve product assortment.

However, agile frontline sales teams, enabled with data-driven platforms, can further nudge retailers to drive additional product lines from the company.

THE PROBLEM

The client has an assortment of over 400 products sold at over 500K retail outlets. Its frontline sales teams often play their conversations with retailers by the ear, and end up missing out on recommending products that can increase revenue lines for the company.

OUR ROLE

We were brought in to enable the sales teams to have a Whitebox approach with the right levers, tools and data to ensure that they were maximizing their product coverage across various retail end points.

Methodology

The Problem was broken down into two simpler problems:

• Which product is a retailer likely to purchase based on its historic behaviour?
• Which products would a retailer willing to experiment with?

Our ML-Based approach was helpful in narrowing down on predictable demand-driven recommendations. The second problem was solved through a combination of recommendation engines that provided recommendations on similar products that would be relevant for each retail end point.

The recommendations are pushed out to a central platform, which then notifies the frontline sales team directly on their handheld devices. The sales teams consumes these recommendations on a daily basis to make the right recommendation to the right retailer.

How did we enable consumption?

• Our solution was first piloted across 6 major regions and then eventually across the nation
• We also diagnosed the deviations and overlaid relevant business context to further tailor these recommendations

Our impact

The Recommendation Engine empowered the frontline sales teams with product recommendations on the fly that could drive more value at the retail end point.

• The recommendations list size was shorter and accurate (accuracy increased by 26.6 %, 800 BPS)
• The recommendation became sharper (recall increased by 14%,1000 BPS)
• The salesman-Retailer interaction time was reduced