Validate Your Own Experiment

The ultimate goal of testing and optimization should not be to simply achieve a lift in revenue or an increase in email addresses acquired.  Instead, the goal of an experiment should be to achieve a learning; to gain an insight into the mind of the donor.

Before we can analyze the results of an experiment, we first need to determine that the results are both valid and that a large enough sample size has been met. This page will help you ensure your experiments meet both of these qualifications.

Validating your experiment is done in two simple steps.

Step 1:

What is the primary goal of this experiment?






Step 2:

To validate the a test for revenue requires entering more than just the total revenue; you will need to enter the individual gifts. You can paste the information necessary below.

Control

Sample Size:
How many visitors saw the page or were delivered the email?

Transactions:
Paste the individual transactions in the field below with each gift on a different line. Pasting from Excel will work.

Treatment #1

Sample Size:
How many visitors saw the page or were delivered the email?

Transactions:
Paste the individual transactions in the field below with each transaction on a different line. Pasting from Excel will work.

+ Add another treatment

Step 2:

Enter the information below to finish your validation. Samples refer to the number of people that saw the test; this could be page visitors, unique opens, or emails delivered. Successes is the number of opens.

Samples Successes Conversion Rate Relative Difference
Control
Treatment #1

+ Add another treatment