Bad site design vs good design

What works best?

[Play_Bach]

Being as it's the new year and all, I decided to put forth what I thought was an improved version of my site. To my eyes, the new site is much better than the old one but my AdSense earnings say otherwise and have gone from bad to worse since launching it. I'm not sure what's going on, so for now, it's back to the old design.

[Lame_Wolf]

Wrong time of year to test things for most sites. Give it a bit longer, say a month before deciding.

[tangor]

If changes are dramatic, and site has recurring visitors, might lose some... site will look different and might believe they arrived in error. "Big Changes" I generally roll out over a 90 day period with small changes every 30 days...

[ember]

People do not like change. If they are familiar with your old design, the new one will throw them and they'll behave differently, which will be reflected in Adsense. Initiate changes slowly and if possible, let your visitors know about them ahead of time.

[explorador]

Don't judge too quickly, a redesign might affect your earnings and traffic in good or bad ways but you have to give it time. The users that come back to your site will need time to get used to depending on your new navigation, also, diff kinds of changes cause diff reactions on the search engines, sometimes they like it, sometimes they don't. Anyway the effect (if good) is not always instantaneous.

One site of mine showed a lot of page views but it was when people were trying to find something, then got back to normal. Other site showed the same but it was derived from more traffic and it stayed that way (traffic went up) but after a while.

[topr8]

... you must decide what you are trying to do.
what is good design? good for who? for what?
do you mean, good: visually, aesthetically, functionally or a bunch of other possibles.

or is good a design that maximises your adsense revenue?

[netmeg]

I generally do a complete redesign off season (for my peak) and I assume that it's going to take Google three to six months to deal with it - everything from from organic SERPS to AdSense targeting. This has worked for me over a number of sites. My first site was redesigned in 2009, and earnings initially went down (probably due to more than redesign) but it bounced back and eventually was earning about three times what it was previously (also probably due to more than redesign)

I for one don't think you can adequately judge the effect on earnings in less than 30 days after relaunch.