Are you interested in changing how Amazon does marketing — moving beyond platform-optimized broad reach to campaigns that find the right customer, at the right moment, using Amazon's unmatched 1P data?
We are seeking an Applied Scientist to join PRIMAS (Prime & Marketing Analytics and Science). In this role, you will design and run the experiments that answer the foundational question for EU marketing: does adding 1P audience signal on top of Value‑Based Optimization (VBO) improve marketing efficiency — and if so, for which customer cohorts, on which surfaces, and at what scale?
Amazon's current marketing model is largely platform‑led: we set objectives and let platforms optimize toward conversion. This approach works well for broad acquisition but systematically underserves lifecycle goals — it cannot distinguish between a Bargain Hunter who will never pay full price and a high‑potential customer one nudge away from becoming a Prime member. This role sits at the center of changing that. You will build the 1P audiences, design the experiments that test them,



and generate the evidence that guides how Amazon allocates hundreds of millions in marketing spend.
Year 1 is an experimentation year. You will deploy 1P audiences across multiple surfaces and channels — Meta, Google, Amazon Display Ads — and measure incrementally against VBO baselines. The goal is not to replace platform optimization but to understand when and where the combination of 1P signal + VBO outperforms VBO alone, and to build the experimental infrastructure that makes this learning scalable.
Key job responsibilities
1P Audience Development & Experimentation

Build and validate 1P audience segments from Amazon behavioral, transactional, and lifecycle data
Design experiments that isolate the incremental effect of 1P audience signal over platform VBO baselines
Deploy audiences across activation surfaces and establish measurement standards that make cross‑surface comparison valid

Causal Measurement & Inc

📌 Machine Learning Scientist (Madrid)
🏢 Amazon Science
📍 Madrid