⎯⎯⎯⎯⎯⎯⎯⎯

Case Study

Business challenge

A multi-national premium condiment manufacturer needed to optimise production for a suite of product with challenging physical characteristics.

With requirements for tight management of uniformity,viscosity, humidity and fineness, and with trade‑offs between efficiency and quality virtually unknown, it was critical that each step of the production process be managed to a very small margin of error.

Business impact

0%+

throughput improvement through process optimisation

0 weeks

to balance linear viscosity without impacting quality

Analysed >200 GB of historical data from 3 sources spanning 4 years to identify the relationships between 350+ production variables

Visualised data to highlight the impact of key variables on yield

Implemented data analytics toolkit to simulate various scenarios and test hypotheses

Defined ideal process configurations with SMEs and ensured required organisational wiring was in place to optimise production yield

Built client team capability to effectively use tools and advanced analytics to further optimise production over time

‟We are impressed by DataStories’ ability to translate data science methodology to a user-friendly platform. The model that results from our dataset are not treated as a ‘black box’ but made interactive and transparent with ‘what-if’ graphs that facilitate conversations with the business and enable us to directly improve process.”

Data Engineer

Key Takeaway

Smart manufacturing programs should be people driven, not technology driven.

Recommendations must be internalized by operators and engineers, else they will not follow them.

Explainability of A.I. is key.

Prescriptive Analytics Left

Consumer Goods and Retail