Comments on: The One Website Test You Probably Never Thought Of… http://www.grokdotcom.com/2007/04/23/the-one-website-test-you-probably-never-thought-of/ Marketing blog focused on marketing optimization, improving website conversion rates, search engine marketing, web analytics, word of mouth, etc. Sat, 21 Nov 2009 23:38:51 -0600 http://wordpress.org/?v=2.8.6 hourly 1 By: Marita http://www.grokdotcom.com/2007/04/23/the-one-website-test-you-probably-never-thought-of/comment-page-1/#comment-716103 Marita Mon, 12 May 2008 18:30:49 +0000 http://www.grokdotcom.com/2007/04/23/the-one-website-test-you-probably-never-thought-of/#comment-716103 I'm ready to change the domain of an existing, but still new site. How do you best do this? Just change it from one day to the next, or do you do a forward? Does it matter? The site gets a few hundred uniques a month; should I worry about losing them, because I'm sure switching the domain is very bad for SEO purposes, isn't it? Thanks for your input :) I’m ready to change the domain of an existing, but still new site. How do you best do this? Just change it from one day to the next, or do you do a forward? Does it matter? The site gets a few hundred uniques a month; should I worry about losing them, because I’m sure switching the domain is very bad for SEO purposes, isn’t it?

Thanks for your input :)

]]> By: Webgeek Paul http://www.grokdotcom.com/2007/04/23/the-one-website-test-you-probably-never-thought-of/comment-page-1/#comment-45165 Webgeek Paul Mon, 07 May 2007 04:46:12 +0000 http://www.grokdotcom.com/2007/04/23/the-one-website-test-you-probably-never-thought-of/#comment-45165 Once while working for a very large online consumer loan acquirer, I ran a test which yielded counter-intuitive results. I added a page to the checkout process at the big front end of the funnel. The page asked simply why people wanted to get a loan today and how they intended to pay it back (all at once, monthly over time, when their tax return showed up, etc). There were drop down responses for both questions. Being the good A-B tester that I am, I sent 80% of our significant volume through the existing flow, 10% through a flow with question page A added near the front, and 10% through a flow with slightly different questions and text input boxes instead of drop downs, question page B, added near the front. The expectation was that the flow would hurt conversion, but that the bank might be able to use the information to discriminate loan losses better down the line, so the net NPV for the flow with increased information might be larger. Much to everyone's surprise, question page A actually increased net conversion by about 8%! A big deal at the fat end of the funnel. Question page B decreased it nearly as much though, a 6% decrease. We immediately went into the usability lab and found out that users thought we were customizing our product offering based on the additional information they had provided, which we believe is what accounted for the increased conversion. We started running different versions of the question page and tried putting the questions in other places in the flow too to try to optimize them, and ended up with a nearly 10% increase in net conversion once optimized. Surprise! Although I still strive to make checkout flow as simple and streamlined as possible, sometime more is more and always testing in production will yield learning. Webgeek Paul Once while working for a very large online consumer loan acquirer, I ran a test which yielded counter-intuitive results. I added a page to the checkout process at the big front end of the funnel. The page asked simply why people wanted to get a loan today and how they intended to pay it back (all at once, monthly over time, when their tax return showed up, etc). There were drop down responses for both questions. Being the good A-B tester that I am, I sent 80% of our significant volume through the existing flow, 10% through a flow with question page A added near the front, and 10% through a flow with slightly different questions and text input boxes instead of drop downs, question page B, added near the front. The expectation was that the flow would hurt conversion, but that the bank might be able to use the information to discriminate loan losses better down the line, so the net NPV for the flow with increased information might be larger. Much to everyone’s surprise, question page A actually increased net conversion by about 8%! A big deal at the fat end of the funnel. Question page B decreased it nearly as much though, a 6% decrease. We immediately went into the usability lab and found out that users thought we were customizing our product offering based on the additional information they had provided, which we believe is what accounted for the increased conversion. We started running different versions of the question page and tried putting the questions in other places in the flow too to try to optimize them, and ended up with a nearly 10% increase in net conversion once optimized. Surprise! Although I still strive to make checkout flow as simple and streamlined as possible, sometime more is more and always testing in production will yield learning. Webgeek Paul

]]> By: Dawud Miracle http://www.grokdotcom.com/2007/04/23/the-one-website-test-you-probably-never-thought-of/comment-page-1/#comment-38143 Dawud Miracle Mon, 23 Apr 2007 12:34:15 +0000 http://www.grokdotcom.com/2007/04/23/the-one-website-test-you-probably-never-thought-of/#comment-38143 Great point, Bryan. I've played around with domain names myself and found that ones that directly to speak to what you do simply seem to perform better. Great point, Bryan. I’ve played around with domain names myself and found that ones that directly to speak to what you do simply seem to perform better.

]]>