Basics of A/B Split Testing For Your Web Site
One of the absolute best ways to continously improve the performance of your web site is by using split testing, also called A/B split testing.
A/B split testing could be compared to a recent eye exam I had... While peering through the machine, the optimetrist asked me "Which image is clearer... A or B?". I would choose between the two and they would make a small change. They would then ask again "A or B?". Each time making a small change. Always taking the better result. This would continue until the results were very sharp. This same "A/B split test" can be applied to your site with similar results of getting you focuses results. For example: Have your web page as it is right now, and call it "Site A". Then make one change to the page but keep everything else the same. Save this to another location and call it "Site B". You then use special software to alternate which page your visitors go to. It keeps switching back and forth, the next visitor sees page A, the next sees page B... The software will do this automatically for you. The result is you will have an even amount of visitors seeing each of the pages. You then track which page lead to more sales. You keep this page and make that your new control, or site A page. The idea is to constantly be doing this. You want to test EVERYTHING on your site, but remember only one thing at a time! Otherwise your results can be impossible to interpret correctly.
Here are some of the things you want to test:
- Headlines
- Subheadlines
- Guarantees
- Your order page
- Testimonials
- Bonuses
- Basically anything and everything!
Some minor changes will make small incremental improvements to your site. These can add up to a significant improvement. Some changes you make will have very big influences. The point is, you are always working towards a positive result.
Tracking The Results
There are a few ways to actually track which page is more successful. Probably the easiest is simply having two separate order forms for each page you are testing. You can then include code on the different pages when they order, subscribe or what ever action it is that you want them to do. By keeping track of how many visitors one order page versus the other, you will see a higher conversion rate in one over the other.
Make It Statistically Significant
In order to ensure that you are getting a proper result, you want to make sure that you have enough visits to show a definite trend. Don't simply put up your A/B test for 30 visitors and think that you have seen the correct results. I run the A/B test with several hundred visitors before I decide to interpret the results.
Free A/B Split Program
Normally the software used for split testing is going to cost you money. I have located a free split A/B program for you to use. You can download it here. It comes with very easy instructions included with it to install it on your site. This program only does the A/B switching for you. You will have to monitor the results on your own by setting up different order pages as explained above.
My Stat Program Of Choice
I use a more robust stastical program on my site. The proanalyzer stats program offers a variety of tracking and statistical results, in addition to advanced A/B split testing. It is a very complete program that I find invaluable. Here is a screen shot from one of the pages you can track your testing on.
You can check out everything it can do for you here (a new window will pop up so you don't loose your place).
Example A/B Test Result
Here is an example screen shot of what a finished A/B test may look like for you.
In this test, a two way split was done. You can actually do a three way split with this program but I prefer to keep it simple. The URL #1 had its sales reflected on the left side of the pie chart in the lighter blue color. URL #2 is showing how successful it was on the right side of the chart in dark blue. If these results were after hundreds of visitors, this is a very good indication that web page #2 was performing better than #1. You should then make #2 your primary page and do another test. Change another small detail and track the results. Work it towards perfection! The bottom line is that some of the best web sites in the world constantly test and track their pages. If you don't you have no idea what is working better for you when you make a change. The sooner you start using Split Testing, the faster your site will be on its way to converting as many visitors into customers for you.
(As with any changes to your site, make sure you test to confirm that everything is working properly once they are put into place!)
| 
About this Site
