Version A has a green button. Version B has a blue one. You might learn something about that execution. You have learned much less about why a customer would choose the product.
Start with a hypothesis about the person buying. Perhaps convenience matters more than price. Perhaps a demonstration is needed because the product is unfamiliar. Perhaps the main obstacle is trust.
Design creative that lets you examine that question. Keep the offer and other important conditions comparable where possible. If you change the message, audience, landing page and discount together, interpreting the result becomes difficult.
Judge against the intended outcome. A curiosity-driven opening might attract clicks while sending poorly matched visitors to the page. A less dramatic creative might lead to more useful purchases.
Write down the result, the limitations and what it suggests testing next. A losing version can still expose a weak assumption. A winning version is more useful when you understand what probably made it work.
