A common misconception of usability testing is that big tests mean big results. While split testing, large scale changes are fine when absolutely needed but they aren't necessary to achieve such results.
Of course minor changes to calls to action can have a significant impact, but other less obvious elements of a page can have a similar impact. For instance, the headline text of a page (typically the <h1> on the page if you're up to speed on SEO) has proven time and time again to impact significantly on the potential conversion rate of a page.
Another example of a minor yet important change is the inclusion, position and selection of images on a page.
Looking deeper into the reasons behind the significance of these changes we come to an ever present usability hurdle - the attention span of users. With the typical user reading at most just 20% of the words on a page in an average visit, (Jakob Nielsen 'How Little Do Users Read?') it's hugely important that your message for the page is clear and concise. Irrelevant images and undescriptive headings are two obvious put offs for visitors to read on, but conveying your message in a way that encourages users to read on is less obvious. As it stands now, no-one other than your site visitors can tell you what they are looking for on each of your pages. So how do you find that out? You guessed it, usability testing.
Usability testing doesn't have to be a case of radically changing a site to see dramatic results, it can be making minor tweaks here and there, the sum of which can result in something pleasantly unexpected.