Difference between events and metrics
- Optimizely Web Experimentation
- Optimizely Personalization
- Optimizely Performance Edge
- Optimizely Feature Experimentation
- Optimizely Full Stack (Legacy)
Events and metrics in Optimizely Experimentation are necessary for experiment and campaign measurement, but they serve different purposes.
Events
Events track specific visitor behaviors, such as clicks, pageviews, form submissions, and purchases. Events are always on, and you can track them in any experiment, not just the experiment where you create them. This makes them reusable and helps you get experiments and campaigns running quickly.
Once tracked, events become the foundation for defining metrics.
Metrics
Metrics measure the impact of your experiments and campaigns by evaluating event data. When you create a metric, you choose one or more events to track and define how to aggregate the data, such as counting total conversions or calculating revenue per visitor.
Metrics help you answer questions like the following:
- Which variation led to more purchases?
- Did form submissions increase?
- Was the bounce rate reduced?
Depending on the behavior you want to measure, you can define the following different types of metrics:
- Unique conversions – Tracks the number of visitors who triggered the event.
- Total conversions – Tracks the total number of times the event occurred.
- Total revenue – Tracks the total amount of revenue generated from the event.
- Total value – Trackstotal numeric value generated per conversion.
- Ratio – Tracks the relationship between two events by dividing one by the other.
For more, see Types of metrics and when to use them.
Key differences between events and metrics
The following table summarizes the key differences between events and metrics:
Events | Metrics |
|---|---|
Track visitor behaviors like clicks, pageviews, or purchases. | Measure experiment outcomes based on event data. |
Collected through SDKs, APIs, or UI configuration. | Created by selecting events and defining how to aggregate them. |
Represent raw data. | Represent calculated insights. |
Must exist before you can create metrics. | Cannot exist without events. |
How events and metrics work together
Events and metrics are closely connected. Events are the data inputs, and metrics are the way you interpret and report on that data. For example, if you create a "Form submitted" event, you can build a metric to count how many times the form was submitted, calculate a submission rate, or measure the total value associated with that action.
The following are example events with their corresponding metric:
- Button click event – Total clicks or conversion rate.
- Purchase event – Revenue per visitor or total purchases.
- Pageview event – Bounce rate or total time on page (using a custom configuration).
Create events
To create events in your project, see the following documentation:
- Create events in Feature ExperimentationCreate events in Feature Experimentation.
- Configure events in Web Experimentation and PersonalizationConfigure events in Web Experimentation and Personalization.
Create metrics
To define and manage metrics for your experiments, use the metric builder.
For more information on metrics, see the following:
- Create and manage metrics in Feature ExperimentationCreate and manage metrics in Feature Experimentation.
- Steps to create an experiment in Web Experimentation or Performance EdgeSteps to create an experiment in Web Experimentation or Performance Edge.