The second part of our article in Ad Exchanger on building a robust experimentation practice in marketing is now live. The key point in the series is two-fold.
- Empirically validating OCI models (like MMM) is a best practice. While advanced econometrics can infer causality in certain cases and for certain models, if you’re making important decisions based on an OCI model, using experimentation to validate it is a best practice.
- Incorporating experimentation into marketing is hard and requires organizational change.
This article focuses more on the need for organizational change and highlights the necessity of building a strong foundation by hiring skilled data scientists, securing executive buy-in to ensure that experimentation is prioritized, and fostering a culture of curiosity where teams feel encouraged to test and learn. Organizations should establish a shared learning agenda with clear objectives for experiments to ensure they produce actionable insights. We also suggest starting with simple tests and leveraging built-in tools from platforms like Google Ads and Meta Ads Manager before scaling up to more rigorous methods such as randomized controlled trials (RCTs).
