Finding causal effects in marketing for KNOWN GLOBAL

The mission: Known’s Strategy team provides cutting edge market research that turns survey data into insights using cutting-edge data science and statistical techniques. Their clients want to answer questions like “does higher brand spend cause better product perceptions?” and my opportunity was to audit and improve their approach to finding these causal relationships. 

How I did it: I worked with this team in an R&D capacity, surveying literature and open source software capabilities for Causal Modeling, a branch of statistics that aims to discover causal relationships between real-life variables. I uncovered a number of potential improvements to their methodology and codebase, leading to tangibly more rigorous results as well as an increased confidence by executive leadership in the robustness of those results.

Testimonial

“Known engaged Janos to help us ensure that our causal modeling pipeline was up to the state of the art in terms of performance and rigor. Through a combination of desk research and experimentation, we worked together to identify the opportunities and risks of each stage in the process, developing a product roadmap that made sense for our clients' needs. In particular, Janos leveled up the knowledge of our team members by bringing well-organized, fair assessments of the literature and the available open source packages available. He made recommendations that took into account our priorities and resources.

“The result is a world-class causal discovery and estimation methodology that we have used to help some of the largest tech companies discover the main upper-funnel drivers of their KPIs. We're confident in the methodology, actionability, and stability of the output, and it continues to be a strong business offering.”

-Brad Deutsch, VP Data Science, Known Global

Previous
Previous

Lacrosse Data Warehouse, BI, and Lead Scoring