paige870•21d agoMost Upvoted

Swapped my old A/B testing approach for Python scripts last month after an ecommerce client in Denver showed me a 40% lift from dynamic pricing.

I used to just run two landing pages and pick a winner, but after their data team walked me through how they automate offers based on session behavior, I ditched manual testing entirely has anyone else hit a ceiling with basic split testing and moved to something more complex?

2 comments

2 Comments

the_rowan21d ago

Oh, I don't know, sometimes simple split testing is still the most honest way to see what people actually want.

julia84321d ago

Yeah I ran into that same wall about six months ago. Basic A/B testing was fine for stuff like button colors or headlines but when you're dealing with user segments that behave completely differently it just doesn't cut it. I switched over to using server side experiments with user cohorts based on time of day and referral source. Night shift workers and morning shoppers are almost like different species. The lift wasn't 40% for me but even 15% on checkout flow was worth the extra setup time.

-1