I thought I'd share an experience with you just to remind you of a simple fact when it comes to "doing stuff" on your website.
I come from the perspective of a vendor. But this is also true for affiliates. The simple fact is that you must measure the effect of the changes you make.
Change a button? Measure it.
Change a headline? Measure it.
Conduct constant split testing to see what works and what does NOT work.
Often, my split tests reveal no significant difference. So the change that I made didn't help me, nor did it hurt.
Other times I strike gold, and sales rise.
THIS time, I totally bombed, and my test slammed my sales down by 45%.
The details of my failed test: I have a landing page that is optimized to get good organic traffic while also appealing to Google's quality score standards. Lots of links to articles, content, etc. But the top of the fold clearly pushes people to click through to my video, which is a VERY effective sales video.
I thought it would be worth testing a splash image from the video, at full size. The idea was to make the image look like an actual video. I put a play button smack dab in the center of the image.
It obviously worked. People clicked on it. The increase in click-thru to my video was 23%. Not too bad at all!
BUT - it's a good thing I tracked sales too. If I only tracked click through, I'd have come to the wrong conclusion.Sales were actually down 45% despite an increase in click-thru.
My theory is that people felt tricked by my "fake video" image. The image was hyperlinked directly to the real video. But the video was on another page. Visitors probably got spooked by this and hit their "back" button in a hurry.
If you want to get an idea of the type of landing page I'm talking about, and the links leading to a sales video, just go look at TruthAboutAbs.com - I modeled my landing page after this.
The moral of the story is TEST and MEASURE. If you make a change to your site that you think is an obviously smart move, you might destroy your sales.
How I test:
I used to use Google Website Optimizer. But now I use Visual Website Optimizer.
Here's a very detailed tutorial on split testing, which explains everything in plenty of detail. From free solutions to paid solutions, it's all there.