We use our expertise to answer this question.
For conversion rate optimization (CRO), a conversion goal is basically the data point that you want to observe.
In a website experiment scenario, you have some sort of idea around “if we make this change to our website, it should cause this other thing to happen.” This “other thing” is the conversion goal.
We work with e-commerce brands, so there’s confusion around the word “conversion.” A lot of times, people think it refers to “order conversion” specifically—like CRO is order conversion—but that’s not the case at all.
“Conversion” is an observable data point; add-to-cart rate, clicks on the navigation, use of search—it’s any user behavior that is trackable on the website. Within the context of a website experiment, a conversion is something that you want to measure to see how the change you’re making is affecting user behavior. Is it improving it? Or is it, at the very least, not making things worse?
For example: You have your test idea and you may not specifically be trying to improve something, but you’re just wanting to make sure that this idea isn’t actually making things worse. And that’s essentially what a conversion goal is.
There are different types of conversion goals from a data standpoint. There are binary goals and range goals. Binary is the far more common type.
About Binary Conversion Goals
Let’s talk about binary goals first; these meet a “yes” or “no” condition and are most commonly used with an A/B testing tool. Did the user perform this activity: Yes or no?
For instance, a simple add-to-cart binary goal conversion would look like this:
- Binary conversion goal: Did the user click the Add to Cart button?
- Condition: Yes or no
- The condition met: Yes, the user clicked on the Add to Cart button.
About Range Conversion Goals
Range goals are where the data can fall within a range of values rather than meet a “yes” or “no” condition. Compared to binary goals, more complicated statistics and math are involved in determining if a range goal is being affected or not.
For example, revenue is a range goal. You can have a purchase that is $10, or a purchase that is $100 or $1,000. But the revenue value does not meet a “yes” or “no” condition. Order placement would meet a “yes” or “no” condition, whereas revenue is a much wider range.
Most testing tools have a concession for tracking revenue (because it’s an important metric), but they don’t track other range goals, like time on page, scroll distance, etc. These are range metrics you typically have to get out of an analytics platform using a data integration between your test tool and your analytics tool.
How to Prioritize Conversion Goals
Here at Omnitail, we prioritize and categorize goals using three classifications: A primary goal, secondary goals, and guardrail goals.
A primary goal is defined in layman’s terms as: This is the major thing that you are trying to observe or affect with your website experiment. There are always two primary goals on any given website experiment that we run: A user behavior goal and a business benefit goal.
- User behavior goal. This focuses on the user behavior that we are specifically trying to affect with this experiment. Did the user do more or less of this specific activity?
- Business benefit goal. What is the consequential business benefit for this experiment? Typically, that’s going to be something at the bottom of the funnel such as order conversion, revenue, average order value, order size—something that has a monetary value directly attached to it.
Secondary goals are behaviors related to the primary goals, but aren’t the experiment’s specific focus. For example, if you’re trying to increase product page use, a secondary goal might be add-to-cart rate, use of search, or something like that.
Guardrail goals are behaviors you want to make sure are not being adversely affected by the experiment, like page load times, website crashes, and unsubscribes.