Senior Data Scientist, Marketing

Cozycozy is the only platform that gives you real-time access to every type of holiday accommodation in one search. Hotels, apartments, Airbnbs, hostels, villas — even treehouses: we aggregate 20M+ listings from 100+ booking sites (Airbnb, , Abritel/VRBO, , and dozens of specialised marketplaces) across 50+ countries and 20 languages, with offices in Paris and Budapest.

Founded by the team behind Liligo and backed by Daphni, Xavier Niel and Thibaud Elzière, we already handle millions of searches monthly and we're growing fast 🚀

Our mission: be the single place travelers need to compare every accommodation option, transparently and in real time, with results suited to their personal preferences.

Our ambition: build the world's best accommodation metasearch and scale it into every market.

This is a deeply technical business with high-stakes steering decisions on paid acquisition : finding the right users, worldwide. Data is its lifeblood: that's where you come in.

About the role

We’re looking for a Marketing Scientist, reporting to the Head of Data, to co-own how we measure Paid Acquisition and help decide where the money goes.

You’ll shape how we deploy a seven-figure paid budget across 50+ markets and platforms (Meta, Google, TikTok, Snapchat). You’ll tackle hard questions with geo experiments, synthetic control and marketing mix models. Did that campaign cause those bookings? Is this market saturated or under-funded? Should we place a 10x bet on one country?

This is a high-ownership applied science role. Your insights will decide which channels we fund and which markets we scale or shut down. You’ll work directly with the founders and CEO, and hand in hand with our Paid Acquisition Lead, who runs the campaigns. You own measurement and the models behind allocation; budgets are set together.

What you’ll own

• Experimentation culture: build rigorous hypothesis testing across the company (difference-in-differences, geo experiments, synthetic control) so learnings are reusable, not anecdotal. Hypotheses often come from Paid Acquisition. You make sure each test can answer its question, and you say no when it can’t.

• Marketing Mix Modeling: own the models behind our paid strategy, starting simple and growing with the business. You decide how we attribute searches to ads and separate paid impact from seasonality and baseline demand.

• Budget allocation: build the evidence and recommendations for spend per market and channel, whether that means shutting down a market, launching a new channel or 10x-ing a campaign.

• Paid Acquisition Hub: build and maintain a lean app that gives a single view of all marketing spend. It lets people or agents act fast, for example by cutting spend in a market when cost per search spikes.

• Data foundations: own the marketing data layer so questions like “How much did we spend on Google in the Netherlands last quarter?” take a minute, not a morning. You’ll work with Carlos, our Senior Data Engineer, who owns the pipelines feeding it.

About you

• Ownership and entrepreneurial spirit: you own outcomes end to end, thrive in ambiguity and happily step well outside your comfort zone.

• Rigor and intellectual honesty: you back every decision with evidence and build setups that can truly confirm or reject a hypothesis. You’re as comfortable saying “this worked” as “we can’t tell from this data.”

• Builder-leader: you raise the level of those around you and challenge the status quo, rebuilding marketing spend allocation with AI and first principles rather than patching broken processes.

• 5+ years in Data Science or quantitative analysis, including 2+ in marketing measurement, growth or a comparable commercial setting. You have strong SQL on a production warehouse and Python for analysis and modelling.

• Hands-on causal inference: you’ve run geo or time-based experiments and can implement diff-in-diff or synthetic control yourself, not just call a library. You run power analyses upfront and will say when a test can’t detect the effect and shouldn’t run.

• Marketing measurement fluency: you understand attribution and its limits, incrementality, and platform-reported data, including why a channel’s own conversions don’t prove incremental effect. You make decisions in CAC, ROAS and LTV.

• Modern stack and strong partnership: you’re comfortable with dbt, cloud storage, orchestration (Airflow/Dagster) and BI (Metabase). You work closely with non-technical stakeholders, the founders and the CEO, and you co-own budget decisions with the Paid Acquisition Lead, grounded in a sharp understanding of a seasonal business.

• Nice-to-haves: experience in travel or large-scale search, Bayesian methods, or having stood up an experimentation platform (randomization, feature flags, a trusted stats engine, warehouse-linked results). You’ll build this capability and champion it across the company.
  

Why join Cozycozy?

🚀 A rare scope: Measurement for the whole paid business is yours. The experiments, the models, the data behind them and the decisions all of these enable.

🌍 Work side by side with the founders and a small, senior data and product team — and build the team around you.

💡 We reward initiative, ownership, and building from the ground up. The platform, models, and automations you ship will steer decisions used across the whole company.

🤖 A mandate to push AI hard — not to talk about it, but to replace real, tedious work with it.

☕ Our Paris (75010) office, with unlimited coffee, tea, snacks, and drinks.

💶 Competitive salary, stock options, Alan health insurance, and Swile card.

Our hiring process

1. First call with our recruiter

2. Technical interview with the Hiring Manager

3. Technical & strategy case, presented to the Paid Acquisition Team and key stakeholders

4. References

5. Offer

What’s Next?

Once we receive your application, we will thoroughly review your CV and cover letter and will return back to you with an update as soon as possible. We invite you to get to know more about us by visiting and follow us on LinkedIn , Instagram, Tiktok and Facebook !