A leading Health Supplement Retailer increases its Revenue from product recommendations by 52% using AI


The client is a leading supplier and retailer of vitamin, protein, sports and herbal supplements in the USA both online and in retail outlets.


  • Customer churn was increasing on the other hand average order value and basket size remained stagnant.
  • The current recommendation engine was rules-based hence recommendations were not personalized to reflect changing consumer behaviour.
  • The client was retailing nutritional and herbal supplement 15000+ SKUs from more than 600 brands through its offline and online channels (It needs a little more explanation about what or how is this challenging).


Tenzai deployed the Purpose Driven AI approach to develop a machine-learning based recommendation engine for the client.

Data from multiple sources like CRM, clickstream data, browsing behavior, past purchases was integrated to create multidimensional microsegments based on demographics, purchase and browsing behaviour. Based on the microsegments different customer genome profiles were defined.

An ensemble of techniques like collaborative filtering, matrix factorization and clustering were used to develop the recommendation engine.

The solution was able to choose the optimal algorithm for each customer based on the context, microsegment and data availability.

The solution has an interface for business users to incorporate business goals and priorities. They could assign priorities and weights to different products based on business objectives.
The multiple solutions were then deployed in docker on the cloud.


The machine-learning based recommendation solution helped the retailer to succeed in its omnichannel strategy.

It helped the company achieve multiple business objectives like enhancing customer engagement, profitability and basket size.

It gave the company the flexibility to prioritize products for a recommendation based on organizational goals.

  • Within the first 6 months of implementing the solution, average basket size (ABS) and average order value (AOV) increased by 20% and 35% respectively.
  • Revenues from recommendations increased by 52% from 17% to 26% of total online sales.
  • Post-deployment, the client was able to reduce the churn rate reduced by 10% resulting in substantial revenue savings