Time to leave the MVP version of https://noisecutters.com behind us and build something nice.
And it brought great results: the sign-up rate immediately went up by 22% (n=300 so far)
Whenever I build an MVP, I don't completely care what it looks like. As long as it doesn't make me cringe, I'm ok with it.
And there's a simple reason for that: I am testing whether it's something that people want. If the design isn't that great, but people still sign up, it is probably something I should dedicate more time to.
I'd rather test something that I put an hour in building, than build something for 10 hours only to find out there's no audience for it.
So, that's what I did.
Here's the MVP design:

Here's the much-improved design:

Old conversion rate: 17,58%
New conversion rate: 21,52% (That's a +22% increase!)
Now, I need to run this change a bit longer. And obviously a full new web design is always something that doesn't really test individual components (which, if you're going to run A/B tests, I much more recommend than just doing a full redesign - but for MVP purposes, this works fine).