Split Testing Ideas for Your Website

By Ryan Deiss

Just about everyone who is serious about digital marketing wants advice about their sales site. In fact, people ask me all the time about what fonts, background colors, or payment options will get the best results on their website.

You see they expect me to know these things…

Frankly, it would make a lot of sense if I DID know the answers to questions like these… The truth is: I DON’T.

So, what do I do?  Just like anyone else, I have to make an educated guess… then watch and wait to see what works and what doesn’t for any given product or niche.

The KEY thing is to build a system for A/B testing all of these variables. And even if they don’t realize it, that’s what most of these marketers are REALLY interested in…

Because using tested sales pages can make you a lot more MONEY.

What is Split-Testing?

Just like it sounds, split testing is basically a testing method that pits two different variables against one another in a showdown featuring live web traffic.

For example, say you were testing which image was more likely to solicit a donation to your wildlife conservation foundation -- a photo of a cute bear cub in a field of daisies, or a picture of a rampaging Grizzly pouncing on a baby lamb.

You’d design two donation pages that were exactly the same, except one version features the cute cub, and the other features the hungry Grizzly.

Whichever version gets a better conversion rate is the winner…

The winner of this particular test may be obvious… BUT you’d be shocked how often the winner is not so obvious. Often, your hypothesis will be DEAD WRONG. Each market is different…

That’s why testing is so important. I can’t stress it enough.

Variables That Whisper Vs. Variables That Scream

Understanding the difference between the two is a MUST for any marketer. Luckily, it’s pretty simple.

Variables that whisper are typically small, stylistic elements that make an impression on a more aesthetic or subconscious level. Here are few examples:

Background colors

Font colors

Headline size/color

Sales letter font

Testing variables like these can be a frustrating and tedious process -- because they don’t produce very dramatic results.  It’s best to save these whispering variables for when you’re getting TONS of traffic and, as a result, you can get a clear and measurable result.

Until then, I recommend you stick with variables that SCREAM. For example:

Price

Headline/Hook

Lead (opening sales message)

Offer (free trial, installments, continuity, etc.)

Variables that scream have a much greater impact on your conversion rates, thus they’re the ideal place to start your testing.

You’re simply not going to see those kinds of results with tweaking a background color. Sorry.

Here’s another key difference: Variables that scream vary wildly from market to market, and product to product. Whispering variables get much more consistent results.

Think about it: Price points in the forex industry just don’t apply to weight loss products.

I can’t tell you what price point will work for your offers, or what type of headlines convert best in your market. Google Analytics and Google Website Optimizer, on the other hand, can help you there.

A Few Variables That Almost Always Win

With that said, I’m going to open up my lab notebook and share a few variables that have won almost every time

These are the background colors, fonts, etc., that have won almost every time in my testing. Once you’ve seen which variables won out for my sales pages, you’ll have a good basis for what may work in your niche… and more importantly WHY it works.

Font Selection/Size – While the typeface you use on your sales page may seem trivial, we’ve seen up to a 30% higher conversion rate by simply using a san serif font. For those of you not familiar with typographer terminology, that means a font without those little “feet” at the bottom.

That means OUT with Times New Roman, and IN with Arial. We’ve found that Arial with a size 12pt font (or larger) gets the best results. Remember, readability is everything. If a prospect has trouble reading your message, you’re screwed.

Background Color – Here’s another variable we get asked about time and time again. For some reason, people are just fascinated with the psychology behind background colors.

We were too, so we tested this variable pretty extensively. Robin’s egg blue (Hex={64,95,ED}) was the clear winner. We actually saw a 31% increase in conversions over dark backgrounds. White and gray backgrounds are also effective.

The theory here is that blue and brown are “trust colors,” i.e. colors that trigger feelings of confidence and, in turn, increase conversions. That’s why so many news agencies use blue and/or brown on their sets.

Table Width – While it’s tempting to revert to the old “bigger is better” theory, our testing shows that 700 pixels is that best size for tables. Make them any wider and your tables lose readability.

In our testing, 700-pixel wide tables improve conversions by 19% over 800-pixel tables. If your market skews older, you may consider a 600-pixel width. Remember, more left-to-right scanning = harder to read = not good.

 

Big Headers – This one’s easy; big headers suck. Even though they may look more polished and professional, our testing has shown that they almost always decrease conversions. They just seem to detract from the headline.

 

Tahoma Red 36pt – This font just seems to work great in headlines, partly because its san serif and partly because it scrunches up more than a font like Arial, fitting more words per line. Red 36pt Tahoma outperforms all other fonts, increasing conversions by 17%. Readability is key.

Drop Shadow – Adding a drop shadow to your red Tahoma font can give you another 5% boost.

Involvement Device – A quiz or poll that’s relevant to your sales message can increase engagement on your sales page. When we tested the marketing quiz below, we got a 24% higher conversion rate.

The key is to keep it relevant and include the words “see results below.” Then, of course, place the results in a space that encourages visitors to read the rest of your sales copy.

Virtual Product Images – These can have a HUGE impact (+145%), if done correctly. Poor product images, however, are worse than none at all. We recommend you hire a professional to do your product images. Poor quality images will make you look bush-league.

Belcher Button – Our own marketing genius Perry Belcher gets credit for this one. He was one of the first to test order buttons extensively and now hundreds of sites use his formula.

As you can see in the image, this ugly little button includes 5 major elements: The orange “Add To Cart” button, red dashed border, marked down retail price, a blue underlined link, and clickable major credit card icons.

Believe it or not, the “Belcher Button” can boost conversions by as much as 300%!

Takeaway

If there’s one thing I hope I’ve drilled into your brain, it’s the importance of using tested designs. There are tons of ways to get this done: Conducting your own split tests, purchasing tested data, or even just evaluating a competitor’s site. But the only REAL way to make consistent improvements to your site -- improvements that make you more MONEY -- is to rely on split tested designs.