Multi-Year Data and AI Transformation for a Global CPG
Three-year enterprise transformation across 24 local + 5 LATAM business units; one of the earliest large-scale enterprise Azure + Databricks deployments worldwide.
Theme Strategy, Vision & Transformation · Also Scaling
In brief
Situation. A global CPG company was operating in one of its top ten markets, a country where it had nearly a century of presence and 99% market penetration, but where data-driven decision-making was still emerging.
Complication. The CEO wanted the country operation to become a beacon of innovation for the wider group. That meant moving from siloed reporting to enterprise-grade data and AI, across 24 local business units and 5 Latin American units, without rupturing operations.
Resolution. I led the technical data team from the original business proposal through three years of execution. The remit integrated data engineering, infrastructure, analysis, experimentation, data hunting, data science, product development, process optimisation, data governance, and software deployment. I worked directly with CEO, COO, VP, and Division Directors to translate business needs into projects. I led teams of engineers, scientists, analysts, digital experts, and product managers, adapting agile, CRISP-DM, and DAMA practices to build a streamlined delivery framework.
Impact. One of the earliest large-scale enterprise deployments of Microsoft Azure and Databricks anywhere in the world. The initiative became a critical component of the company’s digital transformation, recognised both locally and at global HQ. It continues, in evolved form, to this day.
The longer story
Multi-year transformations look impressive on a CV and feel exhausting in real life. The interesting structural insight is that a three-year transformation is not one project; it is roughly twelve projects, each lasting three months, joined by a continuous narrative.
If you try to plan three years in detail at the start, you will be wrong in interesting ways within ninety days. If you try to wing it, you have no narrative and the C-suite loses faith.
The middle path is to plan the narrative at three years and the work at three months. We had a north star, “this company will make decisions with data, and the data will be trustworthy”, and we had a quarterly delivery cadence under it.
The north star changed maybe once. The quarterly plans changed every quarter. That asymmetry is, I think, the heart of how you make long initiatives survive their own duration. People stay motivated by a story. People deliver against a quarter. Confuse the two and you get either drift or paralysis.