Sign Up or Sign up here ?

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.

Loïc d'Anterroches,
Sep 16, 2009

Read more news with tag: A/B Testing, ...