Why most A/B tests fail—and what actually improves conversion rates

Why do most A/B tests fail to produce any meaningful improvement—and what separates a useful experiment from another inconclusive result?

In this episode of Brandformance, Pranav Piyush speaks with Casey Hill, CMO at Do What Works, about what thousands of real-world website experiments reveal about conversion rate optimization.

They explore:

Why most A/B test variants fail to move the needle
The three checks to make before launching an experiment
Why specific messaging consistently beats generic benefits
How reassurance copy can remove conversion barriers
What marketers misunderstand about benchmarks and confidence
Why customer logos alone may no longer build trust
How to position AI products without relying on empty language
Whether websites still matter in a world of AI agents and LLMs

For CMOs, growth leaders and B2B marketers looking to make smarter website and experimentation decisions, this conversation offers practical lessons grounded in actual testing data.