Personalized order recommendation for B2B online sales channel


Client, a Fortune-500 FMCG manufacturer, facilitates the retailers to make orders online through chatbot. However, more than 80% of the sales volume were still happening through offline channels. The client, as part of their digital initiatives, wanted to shift bulk of their orders to online channels instead of offline. Hence, they collaborated with Inxite Out to build personalized order recommendations through chatbot for the retailers.

Personalized order recommendation for B2B online sales channel


Harmonized Data Lake

A centralized data lake is created to translate the data coming from the different sources (E-commerce, Google Analytics / Google Tag Manager, Chatbot etc.) into a harmonized, normalized, unified view which facilitates downstream usage for model development.

Model Development

Utilized factors such as past purchase behavior, short term trends, seasonality etc. to make order forecasts for each retailer-SKU combination on daily basis. These forecasts were further refined by leveraging online behavior data such as recent cart abandonments, purchases in the ongoing week, interval since last purchase etc. to arrive at order recommendations, with added considerations such as stability and control in the supply chain management.


Solution was adopted by the client in the target market across 800+ retailers for 20+ SKUs in the pilot market, with an adoption rate of more than 80%. Success in the pilot led to expansion of this model to 2 more markets.

Case Studies