At the 2025 WPP x Google Cloud AI Hackathon, our team created "Barista" for Nespresso — a real-time budget allocation designed to drive faster and more precise buy decisions for local markets

PROBLEM

Nespresso's static, centrally approved search budgets were limiting growth during high-demand periods and failing to adapt to market nuances, resulting in suboptimal ROAS, missed opportunities and lost revenue.

OBJECTIVE

Develop a scalable, user-friendly Google-integrated tool that forecasts demand, reallocates search budgets proportionally to profitability in near real time, and provides local marketers with clear ROAS and conversion insights

SOLUTION

"BARISTA" combines Google Cloud's Vertex AI, BigQuery ML and Meridian with Nespresso's own data to:

• Predict demand spikes and reallocate budgets dynamically at market and channel level

• Provide and intuitive dashboard for local teams to view CTR, CPC, conversion rate and revenue in near real time

• Integrate seamlessly with existing Nespresso platforms so local markets can act without waiting for central approvals

MY CONTRIBUTION

• Mapped the underlying tech stack and data flow to make it understandable for non-technical stakeholders

• Created and presented a mock example showing how BARISTA would respond to spikes like Paris' Christmas surge or Lyon's underutilised budget

• Framed how different Nespresso teams and agencies would benefit from the tool, turning our prototype into a compelling narrative for judges

OUTCOME

Awarded 1st place out of 30+ teams across multiple cities

TEAMMATES

Dulce María Aguilar (Creative), Ellie Hooey (Strategist), Evan Ozaroff (Data Science), Matthew Reda (Data Science), Ye Tao (Media Analyst), Pavani Pagadala (Data & Analytics), and Saloni Katiyar (Analytics)