Success metric

An experiment is only as valuable as the metrics you track. In the Success metric section of the experiment design panel, you'll define your Primary Metric and any Secondary Metrics. The primary metric is crucial—it determines whether your hypothesis is validated and ultimately whether your experiment is deemed successful or not.

Choosing the right primary metric is essential. If you're new to A/B testing, identifying the most appropriate metric can be challenging, but focusing on the right aspects can significantly improve your chances of success:

• Identify the single user action that best indicates the success of your variants.
• Select a metric directly influenced by the changes made in your variant.
• Ensure the metric fully captures the user behavior you're aiming to impact.

A common mistake is choosing a metric that is too distant from your actual goal. For example, if your variation changes the design or functionality of a landing page, the main metric should focus on user actions directly on that page, rather than a metric like "orders placed," which occurs much later in the user journey and might not reflect the immediate impact of your changes.

Importance of Both Metrics

With Percept Insight Experiment, you have the flexibility to set up several metrics for your experiment. While the primary metric will be the focal point of your analysis, secondary metrics provide a more comprehensive view. For instance, if your primary metric shows a positive result, but secondary metrics indicate negative trends (e.g., increased bounce rates or decreased user satisfaction), it may suggest that the change is beneficial in one area but harmful in others. Conversely, if secondary metrics also show positive results, it strengthens the case for implementing the change across a broader audience.

By considering both primary and secondary metrics, you can make more informed decisions, ensuring that the changes you implement will have a well-rounded positive impact on your users and your business.

Note: Once the experiment is launched, the primary metric cannot be changed. However, you can still adjust or add secondary metrics to refine your analysis as the experiment progresses.