Marketing Analytics & Growth Data · Berlin, Germany
I find the marketing numbers that are wrong, and build the data so they stay right.
I'm Piotr Prędkiewicz, a senior marketing analyst with over 10 years at LOVOO, Trade Republic, komoot and FORMEL SKIN. I've re-valued an audience written off as low-value (ROAS +7%), uncovered ad fraud in traffic that looked healthy (six figures recovered), and built the KPI systems teams actually trust.
About
Measurement people can trust
Most marketing numbers look right. The expensive ones aren't. I work across the whole chain, from the tag in the app to the model that decides where the next budget goes: tracking and server-side tagging, data pipelines and dbt models, attribution and experiments, and the hardest part, getting teams to agree on what a KPI actually means. I like the moment when an accepted number turns out to be wrong, because that's usually where the money is.
Work
Where I've built growth data
+14% AOV
FORMEL SKIN
Built the multi-touch attribution model behind budget decisions, the self-service KPI layer teams ran their business on, and server-side tracking (server-side GTM, Meta CAPI, enhanced conversions).
Creative-level attribution
komoot
Moved user-level attribution to creative-level insight within GDPR constraints; monthly strategic reporting and market research across brand, performance and community channels.
Metrics framework
Trade Republic
Helped define the growth marketing metrics framework and owned the dbt models on Snowflake behind it, including the marketing spend pipeline.
ROAS +7% · six figures recovered
LOVOO
Re-valued a “low-value” audience via indirect LTV (ROAS +7%); SQL checks that caught NULL-tracker ad fraud (six figures recovered); built the marketing data pipeline and automated reporting.
AI in practice
Building with AI agents
AI is part of my daily work, not a talking point. I build my own MCP servers that give assistants like Claude access to data and tools, and run tasks across separate agentic environments. My studio, Solobit Games, runs its marketing on these workflows. I also know when the boring method wins: a forecasting model can beat the baseline on paper and still lose the job in production.
Speaking & writing
Talks, panels & writing
- Roundtable host Data & Analytics: Aligning Your Core Metrics
What happens when every team has different KPIs, dashboards and definitions of success? How organizations align marketing, product, engineering and analytics for clearer decisions and stronger outcomes.
- Roundtable host Did the Click Really Happen?
- Podcast Exposing the myth of 'perfect data'
- Panel Fortifying Your App: Combating Fraud in Mobile Ads
- Blog post Why marketing teams need access to developer resources
Latest project
Solobit Games
An independent mobile game studio building logic-driven puzzle games. Built with AI-assisted production, run alongside my main work. Come see what I'm building next.
FAQ
Common questions
Who is Piotr Prędkiewicz?
A Berlin-based senior marketing analyst and analytics engineer with over 10 years at LOVOO, Trade Republic, komoot and FORMEL SKIN, and the founder of Solobit Games.
What does Piotr Prędkiewicz do?
He makes marketing measurement trustworthy: tracking and server-side tagging, data pipelines, dbt models, attribution and experiments, and KPI frameworks teams agree on.
What results has Piotr Prędkiewicz delivered?
ROAS +7% from re-valuing an audience via indirect LTV, six figures of ad fraud recovered, and +14% average order value from better-measured campaigns.
Does Piotr Prędkiewicz work with AI?
Daily. He builds his own MCP servers and agentic workflows for analysis and automation.
Where is Piotr Prędkiewicz based?
Berlin, Germany. He works in English, German and Polish.
What is Solobit Games?
Solobit Games is Piotr's independent mobile game studio building logic-driven puzzle games. See it at solobitgames.com.
Get in touch
Let's talk data & growth
If your marketing numbers need to be trusted, let's talk.