- Track conversions in your experiment so that you can determine which variation won
- Create and track click goals, pageview goals, and custom events (including revenue)
- Set a primary goal, and learn how primary goals interact with Stats Engine
- Remove goals from an experiment or delete them entirely
Goals at a glance
Goals help you to measure success in an experiment. Use them to track visitor behavior in your variations as clicks, pageviews, engagement, and custom events.
- The primary goal determines an experiment's success. It’s the most important goal of the experiment and decides whether your hypothesis is proven or disproven.
- Goals are saved at the project level and can be used across multiple experiments
- Total revenue is a special type of custom event goal
What to watch out for
Measuring success of your experiment and tracking conversion events is easy with Optimizely. This article will help you set up and manage your experiments’ goals, including:
- Pageview goals
- Click goals
- Custom event goals
- Engagement (the default goal, but we recommend you add more goals that align to your online business goals)
To see more on this topic, check out our two-minute video on setting up goals:
Have you set up a goal, and it's not tracking correctly? Check out our troubleshooting goals article.
Want to learn more about the strategy behind setting up goals in your experiment? Read about defining primary, secondary, and monitoring goals.
Below, we'll cover how to manage your existing conversion goals (you can manage goals both from the editor and from an experiment's results page). Then, we'll discuss how to create new conversion goals.
Already using an analytics platform to track goals? Optimizely allows you to seamlessly integrate with a variety of other analytics tools. Check out this article for more information.
Adding and removing goals
In the editor, click the Goals icon next to the experiment Start/Pause button:
This will bring up a list of the experiment's current goals. When you mouse over a goal, you will be able to edit the goal, remove it from the experiment, or delete it from all experiments that include it.
To create a completely new goal, select Create a New Goal.
It's also easy to share common goals between experiments in a project without having to recreate them for each experiment. To add a goal that you've already created, click Add a Saved Goal. This will bring you to a listing of your saved goals. Mouse over any of the goals in the list and choose Add to add it to the experiment you're editing.
You can remove a goal from your experiment by clicking the Remove button when you hover over the goal on the results page or in the Goals dialog. Removing a goal will just remove the goal from the one experiment. Deleting a goal will remove the goal from ALL experiments and past data collected for that goal will no longer be visible.
Each goal specifies which Platforms the goal works on: Web, iOS, and/or Android. Web goals include Engagement, Clicks, and Pageviews. Mobile goals include Taps, Views, and Custom Events. Custom Events (including Revenue) can be implemented across each of these platforms (as long as they're implemented both on your site and in your app).
We do not recommend adding a goal or variation after you’ve started an experiment. While early on this will unlikely have an effect, as you see more and more traffic there is a higher chance that adding a new goal or variation will impact your existing results. For more information, see our article on false discovery rate control.
When you first create an experiment, change your primary goal from Engagement (the default goal), to the goal that best helps you define "conversions." Most likely, it will be the metric you defined in your hypothesis. For instance, you might want to use:
- Clicks on a key CTA (call to action) on a landing page
- Form submissions on a lead-gen form
- Clicks on video content or social shares, for a media site
- Clicks on "add to cart" or "checkout," for a retail site
Think of your primary goal as the way you'll prove the business value of your experiment.
Why is it important to set a primary goal? First, choosing the right goals for your experiment and organizing a clear report on their outcomes is as important as coming up with effective variations. Second, Optimizely’s Stats Engine treats your primary goal differently from the others for significance calculations.
Optimizely makes sure that the goal you choose as your primary goal always has the highest statistical power, while protecting the integrity of all your goals against inflated error rates from adding multiple goals and variations to your experiment.
All things being equal, your primary goal will reach significance faster than the rest.
You can learn more about the interaction of Optimizely’s Stats Engine and primary goals. We discuss the increased chance of making an incorrect business decision when testing with multiple goals and variations and Optimizely’s solution with false discovery rate control in Stats Engine.
Here's how to set a primary goal:
From the Home page, select an experiment and click Edit next to Goals in the right-sidebar.
This will bring up a list of your Experiment Goals. Drag your goals to reorder them so that your primary goal is listed first.
Types of goals
Pageview Goals: Optimizely records a conversion event for a pageview goal every time a user visits a URL that you choose to track.
Click Event Goals: Optimizely records a conversion event for a click event goal (or simply "click goal") every time a user clicks on a tracked element on your page. More technically, Optimizely binds to the “onmousedown” event to ensure your click goals are captured.
Custom Event Goals: You can use Custom Event goals to track metrics that can’t be easily expressed as a click or URL goal. Some examples of Custom Events are:
- Visitor submits an AJAX form on a page which does not generate a new URL
- Visitor completes a portion of a form
- Tracking clicks on content that is added to the page dynamically
Revenue Goals: You can use the Optimizely Custom Event tracking to track revenue for your experiment.
The Engagement Goal: This goal is measured by default when you create an experiment, and is not recommended to use as your primary goal.
If you encounter any trouble getting your goals to track correctly, read our article on troubleshooting goals for more techniques to troubleshoot your issue.