Experimentation Unite

For many high-involvement products in both B2B and B2C, the decisive steps in the sales cycle often take place offline. As a result, valuable information such as sales stages and customer preferences frequently remains fragmented across isolated silos, leading to ineffective website personalization.
In this presentation, Steffen introduces a concrete architecture for vendor-agnostic data consolidation within a composable CDP. Data from various sources is unified, processed, and activated via interfaces for personalization tools.The result: a complete view for on-site optimization — from the first online journey to the offline deal closure. Learn how connecting these data points enables greater relevance, increases conversion rates, and reduces churn. This setup represents an “opportunity stage” for data-driven growth.

A/B testing has scaled — but impact hasn’t. Many teams declare winners based on secondary KPIs like clicks or add-to-carts, because purchase data is harder to measure and slower to validate.
Over time, proxy metrics become the goal, and experimentation drifts away from what the business actually needs.
In this keynote, Liam explains why most programs over-trust secondary metrics and how this creates “wins” that don’t translate into revenue.
A practical framework is shared for aligning experimentation with business goals: prioritizing the right tests, evaluating most experiments on north star metrics, and validating whether secondary KPIs truly predict primary outcomes.
The takeaway: experimentation that drives measurable growth — not just significant results.

A strategic shift from global dependencies to local autonomy forced us to rethink Kleinanzeigen’s experimentation stack. However, transformation is rarely a straight line.Join us for a candid look at what we got right, what we didn't, and how we're building a self-sufficient experimentation culture from the ground up. We will share why success isn't just about adding a new tool in the organisation, but about taking people along the journey to build a truly self-sufficient, data-driven experimentation culture.