Client Profile: Global hardware giant in Bangalore, with massive supply chain across the globe (2017-18)

Business Results: Savings of around $ 30 million annually, on purchase of various material(OEM/commodity) and with large variation in purchase rate, across and within vendors.

Technical Areas: Data collation, Regression, Decision Tree (didn’t succeed), Random forest(succeeded).

Main Solution steps:

1. Analysing data pivoted on < Given vendor, given product, given data/time range, country, unit price, cost of purchase> to examine the average cost of purchase per unit – min unit cost within six months
2. Predictive aspects: Predicting sale price offered by various suppliers on products against their supplied minimum price /unit.
3. Interventions: Arranging alternatives <confidential>

Client Profile: Large Asian FMCG

Business Results: Turnaround of a failed initiative to control escalatables in decision making.

Technical Areas: Based on categorical division, using “recursive variance adjustment”

— if you haven’t heard of this, don’t worry, we bring the right approach —

Main Solutions steps:

1. A quick review of earlier – failed – attempts (KNN, random forest, SVM)
2. Developed new methodology – classify escalatables, identify data elements, build predictables for decisions of escalation