Building Web Success: Optimizing Online Conversions - Part Three

This is part three of a three-part series exploring necessary steps to optimize your site for online conversions.

In part one of this series, we discussed the importance of setting specific goals for every website and enabling proper tracking to measure and report on web traffic and goal conversions. Part two looked at tagging your goals and measuring your conversion rates.

Completing our discussions with Liam Delahunty of Online Sales, a UK-based company specializing in online conversion optimization, we will now look at how to Test, and Test Again

"Testing the results of minor changes in text and graphics enables you to refine your messaging," states Delahunty. "In most cases, using A/B tests (split tests) or multivariate tests will allow you to quickly see what works best, and why. Finding the message and page layouts that have the best conversion rate will definitely improve your online returns."

As the conversion rate improves, it is important to realize that paid traffic (PPC advertising) also becomes more cost effective. Delahunty offers: "Consider that as you get more sales from the same amount of traffic, you can invest the additional profits back into paid traffic to maximize your efforts. As you continue to test on increased traffic, you will quickly gain insights about your site and visitors - which again, will help in improving the conversion rate. It is an ongoing process." (For more on PPC advertising, visit:http://ppcblog.com)

Some examples of tests your site might use for optimization:

• If a site goal is collecting newsletter subscriptions, offer a free benefit for signing up (tested against no offer) or test using different special offers to see which one has a higher conversion rate.
• If a site goal is to get a specific demographic to call in for product information, create specific and prominent messaging and imagery tailored to them. You might then test different locations on the page to display this messaging, or test different versions of the message itself.
• If your site goal is to increase sales, try testing the effects of minor price changes, discounts or other special offers.
• Test different price points and various bundled product combinations.
• The headline is often key to making the sale. Test various headlines in a multivariate test.

Delahunty warns to not overlook some of the basics. He says to:

• Use the language your visitors use to describe your products or services and not your internal "curse of knowledge" vocabulary.
• Make sure there is a specific call to action for every action you want a visitor to perform.
• Provide specific content on the page for each of the four main personality types that will visit your site:

1.         Spontaneous
2.         Methodical
3.         Competitive
4.         Humanistic

(see: http://www.grokdotcom.com/topics/copywritinghype2.htm for details on these personality types and how they apply to web visitors.)

Don't Try Everything at Once

When you are working on optimizing your conversion rate, Delahunty suggests that you do not change too many things at one time. "If you change many things at once, it can be impossible to get statistically valid results without enormous amounts of visitors." Instead, he suggests making changes that can be isolated and measured against your site's conversion baseline.

"Each small change should be measured, recorded and evaluated," says Delahunty. "When you find something that helps, adopt it and integrate it into your site. If something does not work, note it, log it, and remove it. As more of these tests yield positive results to be implemented, you will see your site's conversion rate rise."

He continues, "Do try bigger changes, too. While there will be a winner established in a test of two similar headlines, testing larger, more dramatic changes will yield results much more quickly. Don't be afraid to see the conversion rate drop, you can learn as much from a decrease in the rate about your visitors as you can from an increase. Just ensure that each change and its results are analyzed and acted upon."

For more information:

On multivariate testing: http://en.wikipedia.org/wiki/Multivariate_testing

On A/B testing: http://adsense.blogspot.com/2006/08/abcs-of-ab-testing.html

On Google Analytics: http://www.google.com/analytics/

On WebTrends: http://www.webtrends.com/

On Omniture: http://www.omniture.com/

On ClickTracks/Lyris: http://www.lyris.com/solutions/lyris-hq/web-analytics/

On Crazy Egg: http://www.crazyegg.com

On Click Tale: http://www.clicktale.com/

On Analog: http://www.analog.cx/

On Awstats: http://awstats.sourceforge.net/

On Webalizer: http://www.mrunix.net/webalizer/

On Liam Delahunty: http://www.onlinesales.co.uk/

SSL Certificate