When doing A/B tests with the color of your button, you are not trying to improve qualitatively (goal of the button, explanation of user motivations).
To me it sounds like you are trying to trigger the subconscious mind faster to take action, so the conscious mind has less time to make a decision. In that case is using A/B tests for determining the color of buttons strictly for maximizing profits and not improving the life of your visitors?
Too me it sounds like you are trying to trigger the subconscious mind faster to take action, so the conscious mind has less time to make a decision.
Isn't that partially a large part of UX? Reducing cognitive load for users and ensuring features are easy to use intuitively?
Why is this unethical? Would you consider Amazon's "One click buy" button unethical because it completes a purchase order faster without the user thinking about it?
Making things easy is not unethical. Making things easy for users to fall into traps / dark patterns or perform actions they do not wish for... now that's unethical.
I'm confused, how is this an ethically loaded question? Persuasion is an essential element of successful product design. Companies need to sell, users need guidance in their browsing and decision making. There's a balance between business goals and user goals that needs to be met. I'd say A/B testing conversion is simply one way to gauge whether your tweaks can make the experience even more effective. This by itself does not mean you go straight into the realm of manipulation. That's what happens if you intentionally deceive your user, make them do things they did not expect or intend.
Here are some true dark UI patterns. You'll see that the intent is to deceive by hiding costs or terms until the last minute, going against common patterns on purpose, or tricking users into letting go of more information than they thought they would be. A button color is none of those things, that's more about visibility.
At my company we often do A/B testing to optimalize designs and conversion rates. Usually they are copywriting and layout alterations, but testing colors would fall under this umbrella as well. Though more specifically we would be testing the contrast of the color scheme (not the color itself). It doesn't matter what color you pick for your call to actions, what matters is how they stand out and work together with the other elements present in the design.
A/B testing for color of your button won't really help you determine the color except if you are looking for an inactive button that would be grey. Me and my team have conducted user research for the company I work for, on exactly the same test and we learnt that it didn't really matter to the user what the color button is. I would recommend that you should go with the color that makes the page/screen consistent.
