Offermatica's Jamie Roche tells us you shouldn't always presume that the most attractive version of your website is the one that will convert visitors to buyers.

This is a real story: A supermodel, who will remain unnamed, works as a spokesmodel for a consumer brand company. The consumer brand creates a website to promote and sell her product using the best practices learned from (literally) years of A/B testing. To put it bluntly, these guys test everything. Following the presentation to the supermodel, she tells the manufacturer that she hates the website that is promoting the product. It isn't pretty, she feels, and it doesn't match her idea of what customers expect from a product associated with her. She has her design person completely rework the site, costing time and money, and she comes up with a more attractive version. The company, which, as I mentioned, then tests just about everything that they do, tests her version of the site versus their original version.

Guess what? The poor supermodel's version gets crushed.

The moral of this story: prettier is not always better -- except, of course, when it is.

I'm not trying to be snide. I'm deadly serious. We have learned, through hundreds of series of testing and optimization, that there is no correlation between how pretty something is and how well it works. In fact, some of your visitors might respond better to pretty while others like you to get them the data as quick as possible.

For example, we recently tested a variety of lovely photographs for the home page of a site that sells bathroom fixtures and accessories. Each big, colorful image was sexier than the last, and we were betting each other on the outcome, backing our favorite photo. The women we polled liked the picture of the model in the tub with bubble bath, candles and a glass of wine. The men liked another one.

It turns out that none of the sexy pictures won.

The most successful home page was the text-heavy, rather bland version. Kind of boring, but man, could it convert. The question is why?

Maybe it's because the shoppers who like this site are plumbers, looking to see if the site has a tub that fits their needs, and they just needed to get in, find the tub, place the order, and go. Maybe for them the image was just a distraction. Maybe the pretty pictures motivated some of their clients, but distracted or put off more of them. It would be interesting to run the test on subsets of the traffic and find out. 

Then, there was PETCO.com. Like most retailers, PETCO assumed that the best way to showcase the products on a category page was to use a shotgun approach: lots of thumbnails on the page to show the breadth of products. Wrong again.

That time, we learned that picking one or two products and showing a larger, more attractive photo converted visitors to buyers at a faster rate than the thumbnails.

Don't be blinded by beauty.

Your site may be as beautiful as Heidi Klum herself, but your conversion rate may still stink. Or, your stunning site has super conversions, and you're chuckling to yourself about how, in your case, pretty does matter. But have you tested your site? Do you know for a fact that the sleek and attractive navigation elements can't be improved?

It all comes back to testing.

Marketing has forever existed in a culture of guesswork based on past experiences. But now that the availability of new technology allows us to double-check our guesses, gut instinct doesn't have to go it alone. Testing and optimization have proven to be valuable assets when executed thoughtfully and thoroughly.

But they can be just as valuable when you don't have the time to give it as much thought. You know what's important to your customers. You know the things that impact their thought process and purchase intent. Even when you don't have time to undertake testing in a big way, test some of those small things.

So don't wait until you finish redesigning your navigation. Don't put it off until next season when you can build a test or two into your timeline. Testing is not a one-time phenomenon. To make it work, to consistently improve the performance of your site, testing needs to be an ongoing process.

Give it a try. You may just find that pretty is better.

Except when it's not.

Jamie Roche is a founder and co-president and CEO of Offermatica. Roche brings to Offermatica the experience of leading a visionary technology company from the dawn of the commercial internet, through the bubble burst and out again. Offermatica, formerly Fort Point Partners, Inc., is an eight-year-old software company that built many of the leading internet commerce websites. With Offermatica, Roche has provided strategic direction to executives from over 50 companies on successful selling through the internet.

Prior to Fort Point Partners, Roche ran Webfactory, a provider of internet products and services to Yahoo!, Netscape and other leading internet companies at their formation. Roche also worked for KPMG Peat Marwick and SiliconGraphics. He is a graduate of Yale University.