Designing Synthetic Controls for Causal Impact: A Medium SEO Case Study

March 7, 2026

This article details an SEO experiment to drive views of my Data Science writings published on Medium. The experiment was inspired by one of my articles, “Demystifying Product Sense in Data Science Interviews”, suddenly earning 200x the average earnings per read, making me wonder how far I could push traffic with a change in SEO setting. Without an A/B test, I used the Synthetic Control Method to evaluate the effectiveness from the Medium SEO metadata update. The 14-day experiment showed a net lift in total article views and demonstrated that a refined SEO setting also brought in a better-quality audience (higher Read Ratio).