In the face of socio-economic volatility, Fast-Moving Consumer Goods (FMCG) businesses need resilience, agility and responsiveness to optimize their supply chains. It means improving efficiency, controlling costs, ensuring availability and maintaining competitiveness.
$30m+
2,100
96%
- 01
The challenge
The business needed to:
- Remove manual supplier analysis to generate enable forecast and decisions.
- Produce fast and accurate responses to rapidly changing pricing, shipping constraints, product availability and storage.
- Save time and cut costs.
- Increase confidence and accuracy across operations.
- 02
The approach
- Using advanced analytics and machine learning, Kubrick developed an Optimizer Engine to model and forecast scenarios.
- Working with Purchasing to outline the variables, limitations, and domain challenges, Kubrick’s data and AI specialists collated largescale datasets from ERP systems, commodity pricing, markets and geo-political data to create a comprehensive model.
- Buyers could test and compare scenarios to optimize purchasing decisions for cost, waste, and longevity of supply and storage requirements.
To facilitate adoption, the team created user guides, demonstration videos, and logs to record and amend any potential data quality issues.
- 03
The impact
The Optimizer Engine enabled buyers to model a cost-optimized purchase in five minutes, reducing decision-making time by 99% and saving 2,100+ hours a year of manual analysis time.
- For the supply of one critical compound, the tool demonstrated annual savings of $2.4m+, with a 96% accuracy on forecasting logistics spend.
- The tool was projected to save up to $30m, targeting six key ingredients accounting for $1.5 billion of spend.
The success of the solution led to a roadmap for a comprehensive Competitive Buying tool suite, providing buyers 24/7 access to insight through a GenAI Virtual Assistant, enhancing supplier risk and performance evaluation, rate card comparison and price forecasting, and resilience to shipping and logistics scenario modelling.
