Skip to main content
All CollectionsRelevic Crash Course
Optimizing Campaigns with A/B and Multivariate Testing
Optimizing Campaigns with A/B and Multivariate Testing
Updated over 3 months ago

Overview

A/B testing and multivariate testing are essential tools in website personalization and optimization. These testing methods help you determine which variations of your content resonate best with your audience by running different versions of a page and measuring their performance. In this article, we’ll explore how to set up and run A/B and multivariate tests using Relevic, along with tips for setting up campaigns and page variations to maximize your results.


1. What is A/B and Multivariate Testing?

A/B Testing

A/B testing involves running two variations of a webpage and dividing traffic between them to determine which one performs better. You might test two different headlines, images, or calls to action (CTAs) to see which generates more conversions, clicks, or engagement. With A/B testing, you can pinpoint small changes that lead to big improvements in user engagement.

  • Example: A clothing retailer might test two versions of a landing page:

    • Version A: Headline – "Shop New Arrivals"

    • Version B: Headline – "Exclusive New Collection"

By dividing the traffic between the two, you can determine which headline leads to more clicks and conversions.

Multivariate Testing

Multivariate testing, on the other hand, allows you to test multiple elements on a page simultaneously. Rather than testing one variable at a time (as with A/B testing), multivariate testing examines various combinations of elements like images, headlines, CTAs, and product descriptions to see which combination works best.

  • Example: A SaaS company might test different combinations of headlines, CTAs, and product images to determine which set of elements drives more sign-ups:

    • Version A: Headline A + Image A + CTA A

    • Version B: Headline A + Image B + CTA B

    • Version C: Headline B + Image A + CTA A


2. Setting Up A/B Testing in Relevic

Step 1: Creating Your Campaign

To begin, you’ll need to create a new personalization campaign in Relevic. Follow these steps:

  1. Navigate to Campaigns: In your Relevic dashboard, go to the "Campaigns" section and click "Create New Campaign."

  2. Enter Campaign Details: Give your campaign a name and set the start and end dates. For A/B testing, this campaign will focus on two variations of the same page.

    • Example: Campaign name: "Homepage A/B Test – October 2024"

    • Set the start and end date for when you’d like the test to run.

Step 2: Setting up Audience Segments

You can target different audience segments for your A/B test by using filters such as Location, Visitor Behavior, or UTM Parameters. This ensures that the right audience is shown each variation.

  • Example: If you’re testing on visitors from the U.S., apply a Location Filter for "United States."

Step 3: Setting up Page Variations

In A/B testing, you’ll need two variations of the same webpage. Here’s how to set up those variations:

  1. Click "Add Page Variation": After defining your target audience, you’ll be prompted to add page variations. You’ll need to enter the URL of the page you want to test.

  2. Create Variation A: This will be your original or control version of the page.

  3. Duplicate and Edit for Variation B: To create a second variation (the "B" version), duplicate the page and make the necessary changes you’d like to test (e.g., a new headline or a different CTA).

    • Example: In Variation B, you might change the headline from "Shop New Arrivals" to "Exclusive New Collection."

Step 4: Defining Traffic Split

After setting up the variations, you can specify how you want to divide traffic between the two versions:

  • 50/50 Split: Divide traffic evenly between Version A and Version B.

  • Custom Split: You can also set a custom split, such as 70% for Version A and 30% for Version B, if you want to favor the original version more heavily while testing a smaller sample on the new version.


3. Setting Up Multivariate Testing in Relevic

Step 1: Creating Your Multivariate Campaign

To set up a multivariate test, the process starts similarly to an A/B test, but with more variations involved:

  1. Create a New Campaign: In the "Campaigns" section, click "Create New Campaign."

  2. Campaign Details: Name your campaign and set the start and end dates.

    • Example: Campaign name: "Multivariate Test – Product Page"

    • Set the start and end dates for your multivariate test.

Step 2: Setting up Audience Segments

As with A/B testing, you can apply filters to target specific audience segments for the multivariate test. Use filters such as Location, Behavior, or Query Parameters to ensure that your variations are only shown to the relevant audience.

Step 3: Creating Multiple Page Variations

Unlike A/B testing, multivariate testing involves testing several variations at once. You’ll create multiple versions of the page that combine different elements:

  1. Add Page Variations: Start by adding the base page (e.g., your product page or landing page).

  2. Create Multiple Versions: Set up several variations, each with different combinations of headlines, images, or CTAs. For example:

    • Variation A: Headline A + Image A + CTA A

    • Variation B: Headline A + Image B + CTA B

    • Variation C: Headline B + Image A + CTA B

  3. Set up Additional Variations: Continue creating additional variations to test different combinations. Relevic allows you to test up to five variations in a multivariate test.

Step 4: Defining Traffic Distribution

In multivariate testing, traffic can be split evenly among all variations, or you can set a custom distribution. For example:

  • Equal Split: Distribute traffic evenly among all variations (e.g., 25% to each of four variations).

  • Custom Distribution: You may choose to allocate more traffic to certain variations based on your hypothesis.


4. Running the Test and Analyzing Results

Once your A/B or multivariate test is set up, it’s time to go live and start collecting data.

Step 1: Publishing the Campaign

Once your page variations are ready, click "Publish" to make your campaign live. Relevic will automatically display the different variations to your selected audience based on your traffic split configuration.

Step 2: Monitoring Performance in Analytics

Relevic provides real-time analytics to help you monitor the performance of each variation. Track key performance metrics like:

  • Conversions: See which variation drives the most conversions (e.g., purchases, sign-ups).

  • Engagement: Compare how much time users spend on each variation, how many clicks your CTAs receive, or how far users scroll down the page.

  • Bounce Rates: Monitor if certain variations lead to higher bounce rates or lower engagement.

Step 3: Determining a Winner

Once your test has run for a sufficient amount of time and gathered enough data, Relevic will help you determine which variation performed the best. You can then choose the winning variation and apply it across your website or continue to iterate and test further changes.


5. Best Practices for A/B and Multivariate Testing

To maximize the effectiveness of your A/B and multivariate tests, follow these best practices:

  • Test One Variable at a Time for A/B Testing: While you may be tempted to test multiple changes at once, it’s best to focus on one key variable (e.g., headline or CTA) in an A/B test to understand its impact clearly.

  • Run Tests Long Enough to Gather Significant Data: Ensure that you run your tests for a sufficient period so that you gather enough data to make an informed decision. Running a test for just a few hours or days might not give you accurate insights.

  • Use Hypotheses for Multivariate Testing: When running multivariate tests, it’s important to hypothesize how different combinations might impact user behavior. For example, will a different image combined with a stronger CTA lead to more conversions? Hypotheses help you stay focused on what you aim to learn from the test.

  • Ensure Traffic is Evenly Distributed: If your test sample sizes are too small or traffic is unevenly distributed, it can skew your results. Make sure that your traffic split settings are aligned with the goals of the test.

  • Track Relevant Metrics: While conversions are often the main goal, track other important metrics such as bounce rate, time spent on page, and click-through rate to get a comprehensive view of how your variations perform.


Conclusion

A/B and multivariate testing are powerful tools for optimizing your website and delivering a more engaging user experience. Relevic simplifies the testing process by providing intuitive tools to set up tests, manage traffic distribution, and analyze performance data. By experimenting with different elements—whether it’s headlines, images, or CTAs—you can gather valuable insights and refine your personalization strategies to increase conversions and engagement.

Start small with A/B testing and progress to multivariate testing as you get more comfortable with personalization. Over time, these tests will provide the data-driven insights you need to make meaningful improvements to your website.

Did this answer your question?