Skip to Main Content

Sep 16, 2008 | 3 minute read

The Magic Buy Button is Just Smoke and Mirrors

written by Linda Bustos

Here's a video that shows a split test - great, we love case studies! YAY!


There's something terribly wrong with the advice given in this video. It encourages you to copy this button design because it was a "winner" in a Google Website Optimizer test - and the premise is this button design will also be the best design for your site.

So what's wrong with taking a proven winning design and putting it to work on your site?

Let us count the ways. For example, what is the color scheme of your site? It's a good idea to test using a button color that contrasts your design. So if you're Office Max, you would want to split test maybe a green, or a blue button against its current yellow-orange design. The narrator mentioned red is a "stop" or "warning" color - but if your color scheme is light blue, perhaps red works best for you. Don't just take this design and assume it converts best for you.

Aside from your website aesthetics, what business are you in? Do you sell electronics? Eco-friendly furniture? How about flowers, candy and gift baskets? If you are a high fashion retailer, does "add to cart" outperform "add to bag?" Would mustard yellow work on a feminine, haute couture site like Couture Candy? Or what if you sell in the UK, does "add to cart" outperform "add to basket?"

You also don't know what other designs were considered losers. The testers likely disregarded red because it is a "stop color." What if red was tested and blew mustard yellow off the hot dog?

Consider this cart button was the top performer for a sales letter for some subscription-based marketing club (for $97 a month!). You know, those sales letters that you're never quite sure what you're buying, but it comes with some kind of free gift valued at $X95? In this case, the "free" gift is DVDs that promise to be a shortcut to actually testing stuff based on some guy's testing:

Don't Waste Time and Lost Sales On Split Testing - Simply Borrow My Results

Here's the big idea. You don't have to waste your time or money split testing. You don't have to spend months and months before you go to market, or thousands of dollars in PPC advertising like we did to learn what really works. You can simply borrow my results.

I can almost guarantee you simply based on the large scientific numbers that we use to gather the results of these tests you could never gather this information even if you had 8 years and a million dollars to waste trying.

Please, please, please don't drink this Kool-Aid. We hope that last week's webinar on testing drove home the value of doing your own testing, and got you excited about getting your hands dirty.

If time and money is a concern, consider that you can at least use Google Website Optimizer - a free tool, and start with a simple test to get your feet wet. The worst that can happen is you learn something new about your customers. At best, you learn something about your customers and improve the profitability of your website.

Tomorrow we'll share a success story of a Get Elastic reader who took the testing plunge and discovered the treatment (test version) realized a 4-6% increase in sales. So check in tomorrow.

For more information and over 250 testing ideas, check out our webinar with Bryan Eisenberg and Bryan's book: Always Be Testing. And check out our webinar replay with Bryan Eisenberg: I Know I Should Be Testing, But...