CRO without the slop
Conversion rate optimization is the most abused word in growth. Here is what good CRO actually looks like.
Most "CRO" is a junior changing button colors and A/B testing copy that doesn't matter.
Real CRO starts with a funnel. Where are users dropping? Where is intent highest but conversion lowest? What does the data say vs. what does the team feel?
A rigorous CRO program runs 5–10 tests a month, each with a clear hypothesis tied to a specific metric. It uses a stats engine (Evan Miller's calculator, or built-in tools) and doesn't call winners until they hit 95% confidence.
It also tests the right things. Hero copy, pricing structure, social proof placement, signup flow friction — these move the needle. Color contrast on the CTA doesn't.
The mistake most agencies make is running tests that are easy to ship, not tests that move the metric. The mistake most in-house teams make is running too few tests to learn anything.
BAZ runs 60+ tests per quarter per client. Most teams run 4. That's the gap.
Keep reading.
Compounding SEO: the only channel that ages like wine
Why SEO is the only growth channel where yesterday's work still pays off next year — and how to build one that compounds.
Read →Paid attribution after iOS: a survival guide
Server-side tracking, conversion APIs, and the new rules of measuring paid media without losing your mind.
Read →How to win the AI Overview
AI Overviews and LLM answers are the new top of the SERP. Here is how to be cited, not just ranked.
Read →