I am a testing fanatic! I firmly believe that there is no way to know if your marketing is performing at its absolute best unless you test. You should be testing your display ads, your sales letters, your email messages and yes, even your website.
I don't seem to have any trouble convincing my clients to test their copywriting and their ads. But, when it comes to website testing, many of my clients shy away. It's easy to look at your site and get totally overwhelmed and unsure of where to start your testing. So I thought I'd share the schedule I plan to use with one of my clients during the upcoming year. What follows is an outline of the website elements I will be testing in 2008. You can use this list to kick off your own split testing campaign.
Remove The Reset Button
The first test I will be performing - because it is one of the easiest - is to remove the reset button from my clients' forms. If you are using a WYSIWYG html editor to create your forms, the reset button is probably generated automatically. And, it's usually placed extremely close to the submit button. Unfortunately, all this does is present your visitor with a distraction and the opportunity to accidentally hit the rest button instead of the submit button. For most forms, there is really no need to have a reset button anyway. In past tests, I've seen this simple change increase conversion rates by as much as 22%. Since it is so easy to create a new page version without the reset button, this is a great place to get started with you're A/B testing.
Make The Submit Button Bigger
After testing the rest button, I'll be moving on to the submit button. Most online marketers agree that when it comes to submit button, bigger is better. So I'll be tested various sizes of graphical buttons until I land on the perfect size. You'll want to keep a close eye on this test as I have a suspicion that a phenomenon similar to "banner blindness" will probably come into play. That is, at some point, the submit button will become so large that it will be overlooked as just another page graphic.
Test The Font Size
Testing your font size is extremely important if your market is older adults or younger children. In the past I have found that a 12-point sans font for body copy converts the best on line. I'd suggest testing that against a 14-point font for kids or seniors. And after you test your body copy, be sure to test your headline.
Reduce the number of columns
The number of columns on your webpage is another easy element to test. Past test data indicates that a one-column page layout performs the best, often yielding over 40% increased conversions. Since my client is currently using a multi-column newspaper page format, I plan to create a test page with only one column of body copy. If your audience is heavily from Asia, I recommend you weight your one column page a 25 or 30% during the testing phase. I've recently come across data that suggests visitors from Asian countries might prefer a more cluttered page layout.
Add dynamic content to your landing page
I'll admit, for the novice this is not an easy element to perform, but since it can have such a huge impact, I am presenting it as an option. One online merchant experienced a 71% increase in revenue simply by adding dynamic data to his landing page. If you're not comfortable with script programming, you might need to get your "tech guy" involved in this one, or search the internet for a ready-made script.
Before you start split testing, be sure you have adequate points of reference to measure against. Study your log files over the next few weeks to determine your current traffic levels, click through rates and conversion rates. This will become your benchmark to gauge all future tests against.
