Retail - Expert Data Science and AI Agents in Performance Marketing

Project overview

Context:

A retail player wishes to accelerate the growth of its international marketplace. The catalog of third-party sellers is about ten times larger than that of own-brand products, but their product pages still generate insufficient traffic. Performance marketing is currently largely manual and fragmented across countries. As part of the 2027 budget preparation, the international team wants to test a centralized approach, based on data science and AI agents, before considering a permanent team.

Challenges:

Assigned Missions:

• Demonstrate that AI can industrialize the performance marketing of an international marketplace.
• Increase the visibility and traffic of third-party sellers' products on AI-based advertising platforms and search engines.
• Manage a growing volume of campaigns and references without a proportional increase in staff.
• Automate simple decisions while maintaining human validation for strategic trade-offs.
• Establish the conditions for international deployment and the creation of a dedicated team.
• Map the performance marketing processes of different countries and identify manual tasks and automatable decisions.
• Audit product data, feeds, campaigns, budgets, and indicators, and identify complementary data useful for decision-making.
• Define priority use cases for AI applied to performance marketing.
• Design and develop AI agents that analyze campaign performance daily and automate certain budget and parameter adjustments.
• Integrate solutions into the existing technological environment: Databricks, Google Cloud Platform, Channable, and advertising tools.
• Optimize product content and feeds for Google Shopping, Meta, TikTok, and AI-based search engines.
• Build a daily recommendation system, with supervision, alerts, and human validation of structuring decisions, in conjunction with the data feed manager and marketing teams.
• Deploy an experiment on a pilot scope, measure performance and productivity gains, then define the conditions for industrialization.

Project details

• Start: flexible
• Remote: hybrid
• Location: Paris, France, FR
• Languages: French