Ever since I have heard about
Distilled’s SEO Split Testing Platform, I have been obsessed with the idea to replicate this tool.
I have read every post I could find about A/B testing engineering, and made this list of resources I found really helpful.
Here are a few things that I learned.
It is hard to collect data properly for SEO Split Testing; The CausalImpact Package is far from perfect. It is hard to detect variations with the package. I will probably migrate to Neural Network instead; If you have to choose between learning R or Python, choose whichever you want, but after a while, I ended up using Python.
You might want to check out Trevor Fox’s post giving
practical tips on SEO split testing.
First, let’s look at the original CausalImpact Package that inspired Distilled’s SEO tool.
SEO Split Testing Resources
Here are a lot of useful articles to learn to make your own SEO Split Testing.
Causal Impact What is SEO Split Testing – Distilled 19 Lessons I learned from a year of SEO split testing CausalImpact Package Documentation CausalImpact Essay Finding the ROI of Title tag changes using Google’s CausalImpact R package Next Era of SEO: A Guide to SEO Split-Testing Compare Actual Vs Predicted Data With Google Analytics And CausalImpact – Bounteous CausalImpact Wrapper for Python Improve CausalImpact Package
If you want to adapt CI package to become more precise you’ll need to investigate the BSTS model.
Research Papers Orthogonal Random Forest for Causal Inference Causal Inference and Stable Learning
SEO Strategist at Tripadvisor, ex- Seek (Melbourne, Australia). Specialized in technical SEO. In a quest to programmatic SEO for large organizations through the use of Python, R and machine learning.