Criteria to use to evaluate response to the A/B test

As part of the planning for an A/B test, you must define the criteria that the system uses to determine the winner.

A/B testing provides three criteria for evaluating responses.

• Maximum unique views. Considers how many different recipients opened the email.
• Maximum unique clicks. Considers how many different recipients clicked at least one link in the email.
• Minimum unique complaints. Considers how many different recipients reported the email as spam or asked to unsubscribe.

Evaluating responses according to unique views indicates how many recipients found the email interesting enough to at least open the email. For example, you might specify Maximum unique views as an evaluation criteria if you were conducting an A/B test to compare email subject lines. When you evaluate according to unique views, the system counts only one message view by each recipient. If an individual recipient opens the email several times, the system does not count any of the views by that individual after the first view.

Evaluating responses according to unique clicks includes recipients that have open the email and clicked at least one link in the document. The system counts any link click in the email, not a specific link. For example, you might use Maximum unique clicks as an evaluation criteria to compare recipient response to different email designs, the placement of an opt-in button, or different formats for hyperlinks. The system counts only the first link click by each individual. Subsequent clicks in the email are not counted in the evaluation results. However, link clicks are also captured in the Deliver system tables. You can view the link click results in Deliver performance reports, such as the Detailed Link and Detailed Link by Cell reports.

Evaluating responses according to unique complaints provides a way to see which message variants your target audience does not like. The system records a complaint when it records an unsubscribe request or an indication from an ISP that the recipient marked the message as an unwanted email, or spam. For example, if you are comparing email subject lines, using Minimum unique complaints as an evaluation criteria can help you to identify subject line variations to avoid for the current mailing. You can use the results of an A/B test that evaluates by unique complaints to improve designs of future mailings. The system counts one complaint for every recipient. It does not count multiple complaints from the same individual recipient.