A De-Siloed Architecture for Product Traceability

• A life-cycle management tool for product formulas
• A life-cycle management tool for product packaging
• An enterprise resource planning (ERP) tool for managing stocks, production, tracking, etc.
• A product information management (PIM) tool, a manually populated marketing tool that collects all product data (photos, description, etc.) and feeds it into the various sales platforms.

Numerous needs requiring the cross-referencing of data emerged: knowing the stock of raw materials required to manufacture a perfume, tracing defective batches so that they are removed from the market, having an overview of the overall consumption of recyclable materials, reducing the high margin of error due to manual population of the PIM (40 hours per month), complying with legal requirements in terms of displaying the list of ingredients per batch, limiting the risks of product incompatibility at a production site, etc.

The Group faced several challenges that made traceability difficult, if not impossible:

• Complicated data cross-referencing, with no established rules or formal processes;
• Unplanned actions, performed ad hoc by the teams as needed;
• Time and energy-consuming processing with data cross-referencing taking up to three weeks;
• An uncertain result because the original data was no longer current at the time.

For these reasons, Cellenza was asked to provide a technical solution to this strategic problem.

Key Results

Implementing a Data Lakehouse-Type Architecture

Cellenza stepped in to implement a data lakehouse-type architecture.

First, Cellenza’s experts implemented a mechanism to replicate data from the various source systems and bring it together in a single location. They then modeled and populated business objects (for example, creating a formula table, a raw materials table, a products table, etc.) and populated any tables that resulted from cross-referencing data from these source systems.

Each source system is run through three layers:

• Copying data from the silo
• Business tables
• The use case, with the cross-referencing of data between the various sources

Cellenza worked on two main use cases: the 360° product vision report and PIM population, including a list of ingredients for each production batch. All the information is available for every product, and the data can be processed in a variety of ways. For example, it is now possible to measure the impact of a marketing campaign on product production.

Once the new architecture was up and running, Cellenza helped the client’s teams improve their skills so they could use the tool independently.

Presentation of Project Results

The impact of implementing this new de-siloed architecture was immediate.

The group can now guarantee the traceability of its products (which was not the case before), and data cross-referencing now only takes a few minutes (compared to several weeks before).

It can also comply with the various regulations in force.

This architecture has also enabled the group to detect incidents in the field and identify new use cases.

Lastly, because the PIM is automatically populated, the risk of data entry errors has been eliminated, and the time saved is tangible (40 hours per month).