Landing Page Testing & Tuning

Apr 11, 2007 • 5:16 pm | comments (0) by twitter Google+ | Filed Under Search Engine Strategies 2007 New York
 

Tim Ash at Site Tuner.com is up first and explains that this is going to go by real fast due to time constraints. What is conversion tuning and why should you care? There is three online marketing activities such as acquisition, conversion, retention. He says the problem at this conference is about getting people to your website. He asks how many spend for on site conversion and testing. He says he wants to help us fix the weak link. The economics of the business is to get cost down on the website. Who should design your website? The real answer is none of the above, it should be designed by the visitor of the website. So what can you tune? You can tune mission critical parts of the website. Landing pages leading to track able actions. You can tune the price of your product or service. What can you tune on your page? What can’t you tune. He recommends don’t go to your competitors website and copy what they are doing. What works for them might not work for you. There are several ways to tune, A-B testing, multivariable testing. With A-B testing you can test one variable at a time. Send equal traffic to all at once. The typical test size is 10-100 recipes. The testing tries to predict best setting for each variable. He explains there testing engine. It uses proprietary math for internet marketing.

What mistakes should I avoid? Ignoring your baseline. He shows some examples of baselines on sites. The conversions improved but the baseline got worse. Always devote some bandwidth to your current version (the baseline). Measure relative to the baseline, not an absolute number. He asks another a question, 90 conversions or 100 conversions. The answer is that we don’t have enough data. Mistake number two is not collecting enough testing. Do not make decisions based on too little data. Pick a confidence level. He nexts puts up some example of pictures of Ferrari and then a Ferrari that had crashed. He explains Volvos are safe! Its not the picture, not the headline it’s the context which you see them in. Interactions are very important. The best setting for variable depends on its content. Mistake number three is ignoring variable interactions. There are major testing themes. Less is more, grab their attention, test the offer, reinforce your key messages, and personalize it. Radical simplification can improve your bottom line.

Next up is Tom Leung from Google. He is going to talk about some common pitfall and various testing concerns. He asks what about if I place the javascript on my pages will it penalize me? There are hundreds of factors search engine consider. Don’t treat search engine bots differently. The variations that you are testing should be consistant. He said don’t be evil with your testing. Javascript tags will not hurt you in the search engines. Testing is still part art and part science so you need to put your thinking cap on or retain a consultant. The Google optimizer works by testing different types of pages. The conversions are recorded and they are telling you what is the best situation and the best conversion.

Previous story: Fun with Dynamic Websites
 

Comments:

No comments.

blog comments powered by Disqus