Skip to main content

    Enhance your marketing with A/B testing

    Course overview
    Lesson
    3 min read

    Analyze your A/B test results

    So, your experiment is complete… Now what? Let’s walk through how to review Klaviyo’s Experiments tab, review findings by channel, and understand what your A/B test results mean.

    Analyze your experiments

    Klaviyo consolidates your A/B test results into 1 place: the Experiments tab. Here, you can click between tabs to review A/B tests specific to each of your marketing channels.

    Note: This tab can only be accessed by Klaviyo users who are account owners, admins, managers, or analysts.

    A snapshot of the Experiments tab, showing recent A/B tests including their: start dates, win probability, status, and more.

    Review results by channel

    For additional information on reviewing your A/B test results outside of the Experiments tab, toggle into each dropdown, specified by channel.

    Campaigns

    To view your campaign test results, you must:

    1. Locate the campaign you tested.
    2. Click on the A/B Test Results tab.

    From there, you can see an overview of the test findings. Plus, review how each variation of your campaign is performing across key performance indicators like clicks and conversions.

    Learn more about analyzing campaign A/B tests.

    Example of the analytics view for A/B test results within the Campaigns tab.

    Sign-up forms

    To monitor a sign-up form A/B test, you must:

    1. Locate the form you tested.
    2. Click on the A/B Test Results tab.

    Here, you can analyze your test results, noting how much revenue, new submissions, and other key metrics are attributed to your test variations.

    Learn more about reviewing sign-up form A/B tests.

    Example form A/B test results, found within the Sign-up forms tab.

    Flow messages

    To view your flow message test results, you must:

    1. Select the flow message that has your A/B test.
    2. In the associated sidebar, click View Test.

    You’ll then be able to review an overview of the test findings and specific data for each message variation.

    Learn more about reviewing flow message A/B tests.

    A visual overivew of the results from an A/B test of an individual flow message.

    Flow pathways

    To monitor your flow pathway test results, you must:

    1. Navigate to your tested flow.
    2. Click Show analytics.
    3. Evaluate the results of each message within your test pathways, noting which saw a better open, click, and conversion rate to determine the winner.

    Learn more about reviewing flow path A/B tests.

    Evaluate your test results

    Upon running your test, there’s 1 key term you must understand: statistical significance.

    Statistical significance is when Klaviyo can mathematically determine whether or not a variation of your content will perform best in an A/B test. It also indicates that you would be able to reproduce the results and could apply what you learned to your future sends. There are 4 categories you will see appear around statistical significance when running your test, as detailed below.

    Statistically significant

    Statistically significant means that a specific test variation is more likely to win over the other option(s). Use this as an indicator of what resonates with subscribers and apply these insights to your future sends.

    Promising

    Promising indicates that a specific test variation appears to perform better than the others; however, the data isn’t strong enough from this test alone to prove a hypothesis. Thus, you will be alerted to re-run the test.

    Not statistically significant

    Not statistically significant means that, while 1 variation has won the test, it has only slightly out-performed the others. As such, the results are not very meaningful. You may choose to test something else going forward and conclude that this element of your marketing did not significantly impact engagement.

    Inconclusive

    Inconclusive signifies that there’s not enough data to determine if a variation is statistically significant. If the test results don’t match any of the criteria above (i.e., statistically significant, promising, or not statistically significant) it will be labeled inconclusive.

    Analyze your A/B test results