Solution
Create Single Source of Truth
- Identified key variables across data sets that could be used to generate recommendations.
- Created pipelines for key interfaces – invoice, product master, customer master, market intelligence, customer spend capacity.
Built ensemble of algorithms to generate recommendations
- Identified patterns in the sales to generate rules.
- Built in business constraints of different verticals and geographies.
- Customer spend capability taken into consideration.
- Generated recommendations across top 5 focus business units and geographies.
Relevancy Check Performance
- Use of multiple attributes from historical sales data to check relevancy of recommendations.
- Validation of recommendations through secondary research.
Continuous model development through iterations
- Cadence with account managers to understand patterns and implementation degree of the recommendations.
- Modifications to the rules of the algorithms.
Outcomes
$540M
of sales opportunity for top 25 customers.
1000
opportunities identified across 4 business units through & patterns.
