Follow

Keep Up to Date with the Most Important News

By pressing the Subscribe button, you confirm that you have read and are agreeing to our Privacy Policy and Terms of Use

Data analysis now gets AI powered precision, Coca-Cola

data analysis data analysis
Case Study by Fractal

data analysisUnifying data with AI-powered precision to yield actionable intelligence 

Introduction 

Lift off into a data harmonization journey 

Organizations worldwide constantly grapple with the complexities of data harmonization —collating and standardizing disparate data sources ranging from unstructured data collected from legacy systems to live-streaming data sets providing real-time information. Bridging the gap between these multiple data sources is paramount for extracting sensible and actionable insights for companies to better serve their existing client base and grow into newer markets. 

Challenge 

Navigating the black hole of unharmonized data 

Our client, Coca-Cola, a household name and a global behemoth in the CPG industry with a valuable brand portfolio and unparalleled distribution capabilities, faced the significant challenge of minutely analyzing data from diverse data sources such as bottlers – separate companies that purchase drink concentrate from The Coca-Cola Company, received from different geographies and in various data formats. Making sense of overwhelming data with no apparent pattern or structure posed several challenges — navigating this intricate data landscape required a precise, adaptable, and detail-oriented approach. 

Processing data manually 

Defining practices revealed a reliance on manual processes, which strained limited resources. Coca-Cola’s evaluation involved manual data harmonization processes, which consumed substantial time and resources. The manual approach presented challenges in conducting detailed data explorations and analyses, influencing product, sales, and marketing decision-making. Additionally, there were difficulties in promptly adapting to market changes and addressing the complexities associated with post-COVID-19 restructuring. 

Managing stakeholder alignment 

Assembling cross-functional teams for data harmonization required aligning all the relevant stakeholders. Coca-Cola’s European data and analytics team, led by stakeholders like the European VP of marketing, VP of sales, CIO, and global data and data engineering leads, established clear expectations, roles, and responsibilities. This challenge focused on ensuring stakeholder alignment and informed decision-making, a critical aspect in navigating the complexities of data harmonization. 

Appropriate platform selection 

Selecting a cost-effective platform presented challenges in evaluating various options. Coca-Cola considered building an in-house solution but opted for an off-the-shelf approach. Key requirements, including NLP engine provision, data stewardship support, and last-mile data integration, guided the evaluation. Harmonizing master data from external companies was challenging because it was incompatible and not owned by Coca-Cola, so they required a solution to address these complexities. 

Solution 

A holistic data analysis and management framework that shines bright. 

The collaboration of Fractal with Coca-Cola outlined the project scope, encompassing the selection of target regions, brand categories, prioritization criteria, identification of data sources, relevant attribute selection, the establishment of standards, and the development of an initial training data set. 

Our Offerings 

An AI-powered data harmonization platform 

Fractal’s customizable platform, Concordia, streamlines data processes with precision and speed. Over four weeks, our team meticulously defined project scopes, including selecting target regions and brand categories, establishing prioritization criteria, and determining data sources. 

We developed a thorough documentation package and an initial training data set. The solution’s implementation involved the development of best practices, constant model retraining, and performance measurement. This agile approach allowed the team to swiftly adapt to market dynamics with well-defined roles across the enterprise and vendors, ensuring seamless collaboration. Deployment occurred through in-house cloud infrastructure, providing flexibility and a systematic training process with clean data sets. This implementation included setup times tailored to each category for higher accuracy. 

 

The approach 

The purpose 

The achievement 

Project scope 

Defining target region, brand categories, and standards 

Comprehensive documentation for project understanding 

Best practices 

Dynamic retraining, testing, and model measurement 

Improved model performance and efficiency over time 

Collaborative roles 

Defined roles and responsibilities for effective collaboration 

Seamless communication and issue resolution 

In-house deployment 

Utilizing Coca-Cola’s cloud infrastructure for flexibility 

Enhanced control, accessibility, and data management 

Training environment 

Clean data-set setup and incremental training process 

Accuracy assurance and improved model effectiveness 

 

Over the moon with global expansion Visible impact 

Concordia’s implementation reduces manual effort and streamlines and automates the data landscape. In addition, scaling up the processing and analysis of crucial data enhances cost efficiency. 

Our client, Coca-Cola, has seen immediate benefits by continuing monitoring practices as the program expands. Leveraging dashboards for data observability will provide a vital tool for assessing and comparing brand performance, offering clear visualizations of insights derived from ongoing data harmonization efforts. 

Reaching for the stars 

In the long term, this solution empowers Coca-Cola to build global partnerships across diverse domains by collaborating with cross-functional partners, including system integrators and consulting allies. This strategy ensures a shared vision and seamless solution expansion, addressing evolving challenges and opportunities for a successful, well-integrated global implementation. Future improvements to the platform by Coca-Cola, such as refined harmonization engines and enhanced integrations, aim to solidify long-term strategic advancements. 

New data acceptance 

Product tracking 

Data lifecycle 

Global expansion 

>96% of new data matched with a corresponding Golden record 

Enhanced unmatched product tracking connects all stakeholders in a regulated way 

Self-sufficiency in data management and insight generation 

New markets targeted by engaging Coca-Cola’s stakeholders and improving the AI platform 

 

Content Courtesy – Fractal