Split Testing Headlines Leads To More Than 122% Increase In Conversion
It's pretty amazing what difference split testing a few words can make. I doubled my conversions!
I took six months off after selling my internet based physical product business in July of 2009. In January of 2010, I started taking inventory of the sites I still owned in an attempt to gather ideas for a new internet based business. One particular site stood out in my mind.
I sell a $30 ebook on this site. I've sold this product for over 5 years and haven't really put a lot of work into it. I don't want to give too many details since I am disclosing the stats below, but I did want to give you an idea of the power of split testing different page elements.
First of all, it is important to note, I've never split tested a single element on this site. This site has done well generating an average of $600 - $800 a month on autopilot, but I was curious to see what would happen if I used a new headline.
I wrote four new headlines. It is important to note, these headlines were not considerably different from the original. Most of them had less than five words changed. One of them didn't have anything changed other than spelling a number versus writing it out. Instead of spelling out million in the headline, I wrote it out like $1,000,000.
I alternated the four new headlines alongside my original headline for 4 weeks. The results were absolutely amazing! Every single headline I wrote increased the conversion over the original by 20%. My best headline beat the original by a whopping 122%.
Take a look at the stats below. The CTA column represents newsletter subscribers and the CTS column represents sales. Clearly title2 is the top performing headline out of the 5 choices.
If this isn't proof that split testing works, I'm not sure what is. For less than an hour of my time setting this up, I've increased my income by more than $800 a month.
There are several pieces of software you can use for split testing. For this test, I used a simple script a friend of mine wrote alongside of Adtrackz. (Adtrackz was not intended for this type of testing, but it works perfect. If you want a copy of Adtrackz, contact me. I'm a reseller. I'm required to sell the software for a minimum of $67.)
Adtrackz is only good for A/B split testing meaning you can't test more than one page element at a time. Multivariate testing is the same principle, but it allows more than one page element to be tested at a time which dramatically speeds up your tests.
Google Website Optimizer offers a free testing tool that works well for A/B or multivariate testing. I plan to put it to use very soon testing different elements on this site.
Have you had success split testing your pages? Let me know if I can help or if you have anything useful to share.
Tags: A/B testing, conversions, multivariate testing, split testing
Would you have a moment to share / upload the test results to A/B Tests?
@Rex – The above results are my A/B test results. I’m in the process now of planning a multivariate test for this site that I plan to start next month. I’ll post those results as soon as they are finished.
@Brent Crouch – thanks for responding, and I have asked @bokardo (Twitter) – the founder of A/B Tests to see if we can post the above or if we need more info.
@Rex – Let me know if you need more info.
@Brent Crouch – Try going to A/B Tests site and seeing if you can post this. Usually it’s a graphic type of change, but you might be able to do the headline type of tests you did. Would be easier for you to do since you have the exact headlines / changes you made.
Just discovered this post. The results are pretty amazing. I wonder how many affiliates are actually doing this (split testing headlines).
Question to you: when you say you’re selling an ebook on your site, do you act as an affiliate of the seller, or as a seller yourself?
Conversion Optimization: Are You Split Testing Headlines? « Affiliates.BlogNotions - Thoughts from Industry Experts Says:
[…] few weeks ago I came across an interesting post that described how the author’s split testing of headlines doubled his conversion rate (and […]
Split-Testing Website Pages – 122% Increase in Conversion Says:
[…] how he had successfully increased the conversions on one of his sites by a whopping 122% by simply split testing […]
Awesome. Split testing is the way to go. We all know it, but few of us really do it.
Franck
Cool results. I started selling a $27 ebook in July. Since then I’ve ran a ton of split tests with about 10 different headlines and also split tested new graphics for the minisite and the conversion rate has gone from a mealy 1% in July to 4% now. 😀 As far as I’m concerned, If you’re not split testing you’re marketing, you’re making a huge mistake!
Thanks for sharing this study Brent, but it seems to me that title 1 performed best since it got the most sales. Please clarify why you say title 2 was best and if you used PPC to drive traffic – thanks.
@Andy – Correct title 1 had more sales, but it received significantly more traffic than the other titles. If you look at the stats for title 1, only .48% of all visitors converted to a sale. Compare that to title 2 and you’ll see that over 1% of all visitors converted.
It was intentional to send the majority of traffic to title 1 during the test. Title 1 was my control. I had used it for a headline for a very long time. I knew how it would perform. I had no idea how the other headlines would perform. They could have done much worse than title 1. So in order to avoid leaving sales on the table, I always send the majority of my traffic to my control during these type of tests.
@Dima – That’s a great increase in conversion. Did you find the headline was responsible for the majority of the increase or were there other on page factors that made a huge difference?