A/B testing is keeping you stuck
Every test starts with a control, and the control is your current thinking. Most teams make small edits from there, which means you're searching for a better answer inside the walls of your existing assumptions. It assumes your positioning is solid, your message works, and you're talking to the right customer. All you need is to figure out which button color closes the gap.
Most of the time, that's not true.
How you get trapped
I've watched brands spend months testing five different headlines that all made the same promise. Every headline was a variation on the same idea. The idea was wrong. No test was going to expose what was possible.
There's another problem with tests that nobody talks about… most a/b tests are local, meaning you're testing one element on one page against a slice of your traffic. A 5% lift there doesn't move your business. It doesn't show up in revenue or anywhere else meaningful.
Why teams don't test the big stuff
The bigger the change, the scarier it is to be wrong. A button color test is safe. Nobody questions why you ran it. If you test a completely different message and it loses, you look like you didn't know what you were doing.
So everyone tests small changes. Small improvements on top of a bad strategy.
Side note… If you want your team to take big swings, you have to reward them for it regardless of outcome.
What actually works
The brands that break out throw out what they had and start over.
They go talk to customers. They find out what almost stopped someone from buying, what finally made them pull the trigger, and what they'd tell a friend. Then they use that language directly in their marketing.
Then they take a real swing. Different positioning. Different price point. Different customer entirely. Most swings miss. But when one lands, it's a 50% improvement across everything, not a 5% lift on a single page with limited traffic.
P.S. Start with a blank page for your next test. Pretend your current site doesn't exist and you just got off the phone with your ten best customers. Build from there.