As you already know, I am doing A/B testing work on the InDefero website. I have first increased by 50% the conversion to get the visitors to access the sign up page. Then, I decided to test the wording of the sign up page. This was just a small experiment to do some A/A testing at the same time.
The results are once again interesting.
I like my daily email with my A/B testing statistics, here are the results for this run:
Test: (2) Sign Up Call
Alt: Sign Up
Conversion rate: 20.28%
Alt: Sign Up (ctl)
Conversion rate: 23.72%
Z-Score: 1.02
Alt: Sign Up Here!
Conversion rate: 23.92%
Z-Score: 1.06
Alt: Sign up here
Conversion rate: 27.08%
Z-Score: 1.90
The first case is the default one with 3 alternatives. I scheduled the test to have 25% on each case. The first alternative is a control alternative. That is, I deliver the exact same content as the control case.
- Good point, the control alternative is not statistically different than the default case.
- Good point, 12 to 25% increase in conversion rate (depending on the use of the default case or alternative as reference).
Why it matters?
If you have 3 steps to convert your visitor and you lose 70% of your visitors at each of them out of 100 visitors you convert: 100x0.3x0.3x0.3 = 2.7 visitors. Yes, only 2.7% even if at each of your 3 steps 30% of your visitors are going through.
Now, imagine you improve by 20% each step, this means that you get 36% of your visitors to go through at each step. It does not look big, but you get 4.67% conversion.
A 20% increase at each of the 3 steps results in a 42% total conversion rate increase! Cumulate small improvements and you win big at the end.
After a lot of bug fixing for InDefero it will be time to rethink the complete sign up process to improve the conversion both for the downloads and for the hosted version.