How To Profit from Googles Not-So-Secret Formula, Part 2

Here is a great write up by my SEO mentor Dan Thies

 

In the first part of this article, we learned how to use Googles Adwords secret formula to improve web site performance, by making your most important action elements more prominent.

 

By doing so, we increased the click-through rate on the links and form buttons that generated the most sales. This improves the web sites conversion rate and profits, just as Googles formula for ranking sponsored listings increases their profits.

 

This time, Im going to give you a couple techniques you can use to drive your sales even higher.

 

Just as Googles advertisers can test several ad variations to maximize their click-through rate (increasing Googles profits in the process), every web site owner can test different variations of their action elements, to optimize their sites performance.

 

There are a couple ways to do this. The first technique is split testing. The second is multivariate testing – a little more complicated, but a lot more powerful, and not that hard to master with the right tools.

 

Ill start with split testing, and a simple example. If youre selling something, you must have a link or a form button that says something like buy now, add to cart, etc.

 

What you may not realize is that the different ways you phrase this call to action can have a dramatic impact on the actions of your visitors, just as subtle variations in an ad headline on Adwords can make a big difference in the click-through rate.

 

In a split test, we simply rotate different variations of the text, and use our analytics software (such as Google Analytics) to determine which version is delivering the best results. Since Google Analytics is free, Ill explain how to do it there.

 

With Google Analytics, you can rotate the different versions and track visitor behavior simply by adding a little Javascript code on the page. The script wraps around the text or image you want to test, randomly selects one variation, and tags the visitor to a segment as described by Google:

http://www.google.com/support/analytics/bin/answer.py?answer=27219

 

If youre split testing the text for a mortgage quote form submission, you could tag visitors with segment names like sendquotenow, sendinstantquote, clickforinstantsavings, etc. based on the text that was displayed, or use generic names like label1, label2, etc.

 

Once youve set the test up, you sit back and let it run until you have a clear winner. You can monitor the results in your Google Analytics reports under Marketing Optimization / Visitor Segment Performance / User-Defined. You can analyze the data based on conversion (for lead generation) or revenue (for e-commerce sites). Use a statistical analysis tool like Splittester.com to decide when the test is over.

 

Stompernet members can watch my Google Analytics Quick & Dirty video in the library, for detailed instructions and downloadable Javascript code. Non-members with a little knowledge of Javascript should be able to work out a solution on their own.

 

Youre probably wondering just how much of a difference split testing can make, right? One of my students recently boosted his shopping cart conversions by over 50% simply by testing different versions of the add to cart and checkout buttons.

 

Thats a huge payoff for a few hours work. These results are not at all unusual in my experience. If you arent testing on your web site right now, its probably safe to say that you are falling behind the competition.

 

Unfortunately, moving action elements around and split testing will only get you so far. Eventually, youll reach a plateau, and thats when youll want to graduate to multivariate testing, using tools like Google Website Optimizer, Kaizen Track, Offermatica, Verster, or Kefta.

 

These tools allow you to test multiple variations in different places on the page. Instead of split testing a headline, you test several headlines in conjunction with several calls to action, several price points, etc. to find the best possible combination.

 

While it is definitely more difficult to set up, the results can be outstanding. In one recent test, a student was able to identify the best combination of text sizes for all the different elements on her product pages, leading to a 9.76% gain in conversions. For a site that we had already optimized with split testing, this was a fantastic improvement.

 

These improvements in conversion arent just a one-time windfall. Every time your testing reveals another winner, it will keep delivering better results for a long time. The game is rigged in your favor – as long as you keep testing, you never really lose.

 

In the third and final article in this series, Ill give you some more thoughts on how to boost your website conversion, including a trick you can use on your site that would be cheating if you tried it on Adwords.