Diageo's campaign launches used to take 15 days. Now they take one.

Diageo's ambition was to scale commerce media campaigns spanning 37 brands and hundreds of SKUs across 60+ markets, where each retailer carries its own specifications, formats, and approval workflows. Multiply the coordination load across all those markets and complexity compounds quickly.

Under the previous model, every SKU variation was built by hand: designers resized bottles, swapped imagery, reformatted layouts, and waited on retailer and legal approvals. Files moved back and forth, revisions were common, and workflows fragmented across tools and agencies.

Nearly a quarter of retail media investment was tied up in production and rework, and as formats multiplied so did the workload. Scaling commerce media didn't just mean running more ads; it meant multiplying manual effort across retailers, markets, and SKUs. The creative ambition was there; the constraint was operational.

“

Shoppers don't browse by category abstraction. They browse by product.

Before Diageo ACE (Automated Consumer Experience), personalization was limited by production capacity. Campaigns were often built at the category level, grouping multiple brands into a single creative like "Save on Whisky" or "Summer Favorites", because generating SKU-specific variations for each retailer was simply too heavy.

DaVinci Commerce introduced an agentic execution layer that restructured how commerce media is built and activated: retailer specifications were codified into the system, brand guidelines were mapped once and embedded into dynamic templates, and product feeds populate the correct bottle, pricing, and messaging automatically.

Each SKU is now assembled into retailer-compliant formats and matched to defined audience cohorts without multiplying manual work, with more than 70% of retailer creative workflows automated. Compliance is enforced in real time rather than corrected after production, so creative teams shifted from resizing and rebuilding to reviewing and refining, while media teams stopped waiting on approvals and started activating campaigns.

Personalization moved from being constrained by production to being enabled by it.

Amazon Prime Day was the first scale test, with seventeen Diageo brands activated simultaneously inside a compressed promotional window. Using modular templates and feed-driven data, the system generated more than 100 Amazon-compliant creative variations in days rather than weeks.

Governance held throughout: retailer specifications were enforced automatically, and campaign launches ran 3.5x faster than historical baselines. And that was just Phase 1.

Following Prime Day validation, the model expanded across additional retail media networks including Walmart and Tesco, supporting onsite placements, offsite programmatic, and digital in-store formats simultaneously. Each retailer's requirements were embedded upstream in the workflow rather than handled as last-mile corrections.

Launch speed improved by up to 15x compared to the prior model, with what previously took roughly 15 days dropping to as little as 1 to 4 days of supervised activation readiness. Commerce media began operating inside the window of shopper intent, not after it had closed.

The performance lift wasn't driven by more spend; it was driven by alignment between timing, personalization, compliance, and execution. Critically, it was repeatable across brands and retail environments, confirming a durable shift rather than a one-off spike.

Shoppers began seeing the exact bottle aligned to their intent rather than a generic category promotion: a premium vodka browser saw premium vodka, and a specific SKU promotion reached the cohort most likely to convert. Granular personalization increased conversion probability, and the combination of relevance and speed compounded performance.

Production costs declined by 76% while return on ad spend doubled versus historical Prime Day baselines, with the savings flowing back into performance-driving media activity instead of operational overhead.