One of the first things we do when optimizing a site is to institute common Best Practices. As important as best practices are, they are still only a baseline for good usability. They are really no more than starting point.
Once you have instituted your best practices and have tracked them in analytics for a while, you now have a place to begin making additional, incremental improvements. This is where multivariate and A/B testing comes into play. Best practices only tell you what works for the masses. Your visitors may react differently, and with some additional and ongoing tests, you may be able improve the overall performance of your site.
There are a number of testing tools that allow you to present different versions of your pages to different groups of visitors. As you make improvements and track the results, you can generate stats on which version had the biggest positive (or negative) impact. With that information, you can make the “winning” changes to your site, which will help you continue to grow your business without having to seek out more traffic.
Don’t let anything on your site be so sacred that you won’t test its performance against another variant. If there is a good idea–or the possibility of a good idea–test it to see if it truly is. It’s not about proving anything right or wrong, but simply about finding what works best for your site and visitors.
Change for the sake of change isn’t valuable. But changes made because they have proven to improve your sales, profit, ROI, etc., are worth seeking out. If you’re confident that you’re doing the right thing, test it. If you’re wrong, you still win because you now know what does and doesn’t work.
Always keep improving your website. Never stop testing. Your visitors and your profits will thank you!